Latest revision as of 15:01, 19 April 2018

Understanding and minimising our digital shadows

The internet is a great space to explore, learn, speak up, listen and communicate with people across the world. Unfortunately, the internet has also become a space where people who challenge the dominant discourse often find themselves under attack. These attacks can be very personal - enabled by the fact that there is often a lot of personal information about us on the internet.

To strengthen our defences against these kinds of attacks, it’s a good idea to start by our assessing our digital shadows. These shadows - can tell a story about us: who we are, where we live and hang out, what we are interested in, and who our friends and colleagues are.

This can expose us to several threats. In particular, it is the publicly available traces we leave behind that expose us to online harassment.

However, there are also many strategies and tools we can use to shape or control our digital shadows, to increase our privacy, and ultimately to be more secure, both online and offline - without being less vocal or reducing our activity online.

Some examples of these include controlling the amount of data we give away by consciously stripping valuable information from content and metadata; trying the art of self doxxing; and thinking about ways to play with and break up our online identity.

What is a digital shadow?

Our digital shadows can be defined as the stories data tells about us. These digital shadows are created by trillions of bits of data, digital traces we leave everyday when we connect to the Internet, our mobile phone and online services. Our digital shadows have a life of their own, are affected by others and change in unpredictable ways. Our digital shadows grow continuously, can be permanent and we have little control over them. These traces are a spectre of our past and present activities, which melt together in a permanent and ever-changing profile.

How are these trillion bits of data created? The devices and the software we use to browse the Internet, access websites, connect to social networking platforms like Facebook and Twitter, publish blog posts, receive phone calls, send SMS messages or emails, chat, or buy things online, all create specific bits of data about us. These bits of data can include our name, location, contacts, pictures, messages, tweets and likes, but also the brand of our computer, length of our phone calls and information about which websites we visit. These data traces can be put out there by ourselves as well as other people.

How do we share data? In some cases we actively share data – for example when we share photos on Facebook, book a flight ticket online or contribute to a wiki. Other people can also actively share data about us, by tagging us in pictures, mentioning us in tweets or simply by communicating with us. In other cases, we give away data without necessarily realising it, or consenting to it. Our browsing habits and IP address are shared when we visit a website by means of "cookies" and other tracking technologies, which are active in the background. These technologies are embedded in the websites we visit, and the information shared is collected for a wide range of purposes, from website analytics to advertising. Our mobile phone apps also collect data on us without our active knowledge or consent – for example, the photos we take usually have location data embedded in them. These tracking technologies enable web services to identify and follow us as we move from one service to another - from our internet browser to the IM (instant messaging) app in our smartphone, from downloading e-books in our readers to publishing pictures from the latest protest we have covered.

What is data? Data can be broken into three parts: content, metadata and noise. Content is the content of our messages, blogs, tweets and phone calls; it is our pictures and videos. Metadata is data about data, information that is needed for the technological infrastructure to work. Metadata enables our email to be delivered, help find files on our computer and permit mobile communication. Metadata can be our email address, phone number, location, time and date when a message was sent or stored. Noise is the data that is created by either the manufacturing process or by the workings of the infrastructure. For example, every camera has an SD card to record and store pictures. Every SD card has unique scratches that were created by the machines producing the SD cards. These scratches make small changes to the data that are not visible to the eye but can be recognized by computers.

Who is collecting our data, and why?

We might wonder about the importance of one picture or one message, or think there is so much data out there that nobody knows what to do with it. However, data collection and data analysis has become very sophisticated.

The data traces you leave are collected, analysed and sorted by various parties to create digital shadows, or profiles. Every time a new piece of data is collected, it can be identified and added to your profile. These profiles are ever-expanding, and give those who create them or who have access to them an immense insight into who you are.

Data is collected by companies, governments and individuals for a variety of purposes. It can be bought and sold; it can be used to control; or it can be used to create harassment strategies.

Our digital shadows or profiles can be used to gain insight into who we are, what we do and where we have been. This data can then be used to make predictions on what we might do or where we might be in future. For example, if someone knows that we are an outspoken blogger on gender issues in country x, they know that we will probably be present at a conference on blogging and women held in that country. Profiles can also give potential harassers the ability to harass us across different platforms.

Anyone could potentially have access to our digital shadow – including communications service providers, law enforcement agencies and commercial companies, as well as groups and individuals running their own servers. We can't know exactly what is happening to our digital shadow, and that itself is a problem.

However, there are tools and tactics we can use to manage our digital shadows and to limit their ramifications in terms of profiling and surveillance. This will be discussed in the rest of this section.

Exploring our own digital shadow

As we mentioned before, anyone can potentially access our digital shadow – including communications service providers, law enforcement agencies and companies, as well as groups and individuals running their own servers. We cannot know precisely what happens to our digital shadow and that itself is a problem. But there are tools and tactics to manage our digital shadow and to limit its ramifications in terms of profiling and surveillance.

Some good places to start are:

• Exploring our individual digital shadow with Trace My Shadow – https://myshadow.org/trace-my-shadow – a tool launched by Tactical Tech, accompanied by a website that offers a lot of tips on how to protect our privacy and control our digital shadow: https://myshadow.org
• Identifying and materialising social networks across our online and physical activities: John Fass, researcher and designer at the Royal College of Art, has created some activities to materialise our social networks and browser history ' [insert link].'
• Seeing through the eyes of our mobile phone by installing a tool called openpaths.cc. Some of our apps can see the same things. Read the Terms of Service carefully and explore if you can change the access settings in your phone. On an iPhone we can change the permissions for each app under its privacy setting.

Control the content and metadata you share

The good news is that we can partly control what content and metadata we give away.

When we publish content on the web, it is always a good idea to ask ourselves if what we are posting is public or personal and who could have access to it. Even if the information is connected to a public event and not to our personal lives, the names we mention or the images we upload may contribute to a picture about who we are, what we are doing, where we are doing it and so on. This could be used by people who wish to target us.

This does not mean that we should silence ourselves – by taking some basic measures, we can limit our risks by increasing the level of the effort that would be required to attack us or our contacts.

• When sharing personal details about our life, we can use private profiles that can only be accessed by selected contacts. When using private profiles on commercial social media, we should be aware of the regular changes to the privacy policies of that platform. This can have an impact on how “private” our profiles are. There have been cases where privacy settings have been changed, exposing pictures, content and the conversations of private groups.

• When writing or posting images about public events on the web and on publicly accessible social network profiles, we should ask ourselves if the information we spread about single individuals, places and other details could be used to identify and/or attack someone. It is always a good idea to ask for permission to write about individuals and perhaps also to post information on public events only after they are finished.

• Faces in pictures can be anonymised with a tool called ObscuraCam, a free camera application for Android devices. https://guardianproject.info/apps/obscuracam

• When giving personal information to a web service, it’s best to use HTTPS so that the communication channel is secure (see the section on security measures for more on this).

• Use strong passwords, and use different passwords for each web service you use - if you you the same password for multiple services and someone intercepts your password for one of these services, they could use it to access your other accounts.

• When registering a device or software such as Microsoft Office, Libre Office, Adobe Acrobat and others, not usingy our real name can help prevent the metadata created when using this device or software from being connected to you. You can also switch off the GPS tracker in your phone or camera.

• Some file types contain more metadata than others, so when publishing contents online you can change files from ones that contain a lot of metadata (such as .doc and .jpeg) to ones that don’t (such as .txt and .png), or we can use plain text.

• Remove metadata from image files by using Metanull for Windows: https://securityinabox.org/en/lgbti-africa/metanull/windows

• For editing or removing hidden data from PDF files, Windows or MAC OS users can use programs such as Adobe Acrobat XI Pro (for which a trial version is available). GNU/Linux users can use PDF MOD, a free and open source tool. However, it doesn’t remove the creation or modification timestamp, and it also doesn’t remove the information about the type of device used to create the PDF. 

• For more on removing metadata from different file formats, see Tactical Tech's https://securityinabox.org/en/lgbti-mena/remove-metadata.

• You can prevent the tracking and collection of metadata through your browser by installing add-ons like Privacy Badger or Adblock Plus, as well as by monitoring our privacy settings and deleting cookies on a regular basis.

• Using Tor will hide specific metadata like our IP address, thereby increasing our anonymity online.

Social domains

As security expert Bruce Schneier explains, “Security is a chain, and a single weak link can break the entire system”.

Each of us belongs to several social domains - our work or advocacy networks, our family networks, friends, and sports teams.

Some of these networks may be more secure than others. For example, we may tend to have a more secure communication practices for our work or advocacy activities, but less secure practices for interacting with friends on a social network. Areas where these domains intersect can turn into a threat to our security.

If we use a single identity in all our domains, it becomes easier to gather information about us and to identify our vulnerabilities. For example, if we reveal in a social network that we like a particular kind of game and that we download files with a p2p program like Emule, an attacker who wants to investigate our work or advocacy activities might trick us into downloading a game which is infected with spyware.

This attacker is interested in our advocacy activities, but knows that we have increased privacy and security measures for that part of our lives. The attacker also knows that our love of games is a digital weak spot as this network is not encrypted. Thus the attacker can exploit this part of our lives to gain access to another, more secure part.

This is only possible, however, if our work identity and our Emule profile can be connected to the same person; and this is why separating our social domains can be useful. More on how to do this will be addressed later on, when we talk about identity management.

Self-Doxing

Doxing (also written as "doxxing", or "D0xing", a word derived from "Documents", or "Docx") describes tracing or gathering information about someone using sources that are freely available on the internet. This method depends on the ability of the attacker to recognise valuable information about their target, and to use this information for their own ends.

Doxing is premised on the idea that "The more you know about your target, the easier it will be to find his or her flaws”.

Harassers and stalkers use several tools and techniques to gather information about their targets, but since these tools and techniques are mostly public and easy to use, we can also use them ourselves. "Self-doxing" ourselves can help us to make informed decisions about what we share online, and how. Of course, these same instruments can also be used to learn more than is immediately obvious about someone we have met online before we give them our full trust.

Methods used for doxing include exploring archives, yellow pages, phone directories and other publicly available information; querying common search engines like Google or DuckDuckGo (https://duckduckgo.com); looking for a person's profile in specific services; searching for information in public forums and mailing lists; or looking for images that the target has shared (and for instance may have also published in another, more personal, account). But it can also simply consist in looking up the public information on the owner of a website, through a simple "whois search" (see the section on "Creating a site of one’s own").

We can use these same tools to explore what can be easily found out about us by others.

Before we start exploring these web services and looking for our digital self, a good idea is to use anonymisation tools like Torbrowser.

• Useful tips on how to self-dox are available here
• For more about (self-)doxing tools and techniques, visit: https://modelviewculture.com/pieces/investigation-online-gathering-information-to-assess-risk
• A useful (and creepy!) tool to learn what traces we have left behind in our Facebook account is Ubisoft’s Digital Shadow, a Facebook app which illustrates what third parties can know about us through our Facebook profiles: https://digitalshadow.com

Can we remove our digital past? Creating new identities

"Once something is on the internet it will stay on the internet, as the internet does not forget".

We may think that deleting certain sensitive data from social networks and web services may be enough to protect ourselves, but metadata cannot be deleted as easily. And using just one identity through our whole life - in all our work and personal domains - creates a bulk of information that could be used to profile or attack us.

One option to avoid this is to leave an old identity behind and create a new one or several new ones - one for each of our social domains. We might also choose to use our real identity in some areas, and our new alternative identities in others.

• When we create a new identity, we should select the contacts for each one carefully, and avoid sharing contacts with other identities we use for different activities. This effectively creates separate social domains, with separate accounts, mail addresses, browser profiles, apps, and possibly even devices.

• It's important to make sure that our various identities are not linked in any way to each other, or to our real identity. Remember that some of these connections can be tenuous: for example, did you sign up for a new, pseudonymous Gmail account using your real phone number?

• Treating each of our extra identities as potentially disposable can be useful, as they can be discarded easily if it is compromised.

• Disposable identities can be created for new acquaintances where appropriate – introductory profiles we can use to get to know somebody before we include them in a more trusted network.

To learn more about how to separate different identities into separate profiles, read the section on “Managing multiple online identities”.

Deleting identities

If we decide to separate our domains by creating multiple identities, one of the first decisions we should make is whether to delete or keep the identity or identities that we already have.

To do this, we can start by investigating the traces of our existing identity or identities. (for methods and tools for following your own digital traces, see "Exploring your digital shadow" and "Self-Doxing").

If we want to delete existing accounts:

• The Suicide Machine (http://suicidemachine.org/) is a tool that facilitate the process of deleting social network profiles. The Suicide Machine was forced to stop deleting Facebook accounts, but instructions on how to do this are here: https://www.facebook.com/help/224562897555674
• AccountKiller (https://www.accountkiller.com) has instructions on how to remove accounts or public profiles on most popular websites.
• JustDelete Me (http://justdelete.me) is a directory of direct links to delete accounts from web services.

Mapping our social domains

To separate our social domains, it's helpful to first map them out and identify which ones could expose us most to cross-domain attacks.

We can do this by thinking about our different activities and networks, and reflecting how sensitive each of these is. This will enable us to better separate the domains that are sensitive from those that are not.

Partitioning one’s digital life into separate social (or "security") domains requires some thinking.

Polish computer security researcher Joanna Rutkowska has worked extensively on this, to the point that she developed  a security-oriented Linux distribution based on the concept of “security by isolation” (called Qubes OS). In this system, each social domain is isolated in a separate virtual machine. While Rutkowska's scheme is quite sophisticated and focused on her operating system, it can give us interesting insights on how to start thinking about separating our domains.

The three basic domains Rutkowska identifies for herself are “work”, “personal”, and “red” (the untrusted, insecure area).

• The work domain includes her work email, where she keeps her work PGP keys, where she prepares reports, slides, papers, etc. She also has a less-trusted “work-pub” domain for things like accepting LinkedIn invites or downloading pictures for her presentations. To add to this, she has a “work-admin” and a “work-blog” domain, in order to get a further level of security for managing her company’s servers and for writing on her blog.

• The personal domain includes all the non-work-related stuff - such as personal email and calendar, holiday photos, videos, etc. She adds to this with a special domain called “very-personal”, which she uses for the communication with her partner when she is away from home. The couple uses encrypted mails to communicate, and she has separate PGP keys for this purpose.

• The red domain, on the other hand, is totally untrusted. This is where her disposable identities or profiles belong. Rutkowska uses this domain to do everything that doesn’t fit into other domains, and which doesn’t require her to provide any sensitive information.

• Besides these three main domains, Rutkowska has several other separate domains. One is dedicated to shopping, for accessing e-commerce sites. What defines this domain is access to her credit card numbers and her personal address (for shipping). Then there is the vault domain, the ultimately trusted place where she generates and keeps all her passwords (using KeePassX) and master GPG keys. Finally, she has a domain for all the Qubes development ("qubes-dev"), one for accounting, and another one for work archives.

Of course we don’t have to separate our domains in such a complex way, and using Qubes Os to keep them separated is just one solution – and one that requires a powerful machine to run on. Yet Rutkowska’s reflections on domain mapping can be an enlightening starting point to analyse our activities, and to separate our social domains for enhanced security.

• Joanna Rutkowska’s article on security domains can be found here: http://blog.invisiblethings.org/2011/03/13/partitioning-my-digital-life-into.html

2. Assessing risks and potentials: how to choose which online identity fits our purpose

"Real" or virtual identity?

Once we have identified our different social domains and the digital activities and contacts that go with them, what we need to do is decide if we want to differentiate our identities accordingly, or if we'd rather stick to our official name and true face for each of them.

We may want to keep our work connected to our legal or "real" identity, or think that our activism should be anonymous, but these are decisions that need to be thought about carefully.

For example, a journalist who finds it convenient to use her real identity for her writing may decide to stay in contact with her personal domain through a nickname, so that nobody can connect the two spheres together.

On the other hand, if an activist decides that she wants to use a pseudonym for her online activities, she should consider that she will be showing her face in all her connected activities in the real world, such as speaking at conferences or participating in demonstrations. Her online pseudonym will therefore be linked to her face; but her face could also be linked to her real name on social networks, and her online activist identity unmasked.

In assessing which identity to use in a given context, it's helpful to consider the following questions:

• Would my job, livelihood or safety be at risk if my real identity were known in this context?

• Would my mental health or stability be affected if my participation in X were known?

• Would my family or other loved ones be harmed in any way if my real identity became known in this context?

• Am I able and willing to maintain separate identities safely?

Once we have assessed our risk, we can then consider different strategies for separating our identities online. 

For more on assessing risk, visit: https://ssd.eff.org/en/module/introduction-threat-modeling  

Strategies for separating identities online

strategies for maintaining separate identities can range range full transparency to full anonymity.

Author Kate Harding talks about her decision to start writing under her real name, dismissing the recommendations that are generally given to bloggers to follow practices like “writing under a pseudonym, making that pseudonym male or gender-neutral if you’re one of them lady bloggers... masking one’s personal information, being circumspect about publishing identifying details, and not writing anything that might inflame the crazies”.

Instead of putting responsibility on women, Harding says, problems of harassment should be handled by society as a whole, including men. However, she also acknowledges that the decision is dangerous one. http://kateharding.net/2007/04/14/on-being-a-no-name-blogger-using-her-real-name

In "The Girl's Guide to Staying Safe Online," Sady Doyle writes that while becoming visible "creates a specific vulnerability", giving up on our online activities is exactly what the misogynists and harassers expect from us - and so the best way to ‘stay safe’ online may simply be to stay online. "After all: If there’s no one left willing to complain about the harassment, what are the odds that it’s going to change?” http://inthesetimes.com/article/12311/the_girls_guide_to_staying_safe_online

As such, we have several options available to us: We can attempt to participate online anonymously. We can use a "persistent pseudonym".  We can identify fully with our real or legal name, or we can divide our online lives, using our real or legal name sometimes and a pseudonym at other times. The following section will explore the pros and cons of each option, so we can determine which options are best for each one of us.

Anonymity

Total anonymity can be both isolating and difficult to maintain, but is useful in settings where we don't need to gain other people's trust, when there are few or no people we can trust, or when we don't want to expose others in our life to risks.

Anonymity may be a good choice in certain specific situations, such as researching or participating in message boards about health issues, or when sharing sensitive information. We may wish to set up a one-time account, using a pseudonym, to comment on a blog or news site, or a one-time email account or chat session to discuss sensitive information with others.

On anonymity, Vani, a human rights activist, writes: “I am a regular social network user. I voice my opinions on a range of topics. But I remain faceless and nameless”. speaks of their participation on social networks anonymously

But while anonymity is a good option in many situations, it can also be dangerous in some countries, where it can signal to the state police that the author thinks they are doing something wrong. This strategy can also be lonely: “Anonymity also isolates you”, a blogger writes. “Can you have a network to protect you and also be anonymous at the same time? Would visibility be a better strategy for you?”

Anonymity differs from "persistent pseudonymity". When we adopt anonymity as a strategy we may use pseudonyms, but these pseudonyms are not used across different networks or social domains, and some may only be used once and then discarded. For more information on how to be anonymous online, see Anonymizing tools.

Persistent Pseudonymity

Persistent pseudonymity involves using a pseudonym consistently over a period of time.

The Oxford English Dictionary defines a pseudonym as "a fictitious name, especially one used by an author." In the age of the internet, a pseudonym may also be referred to as a "nickname" or "handle", though the latter can also be tied to a person's legal identity. 

Jillian C. York, in a blog post for the Electronic Frontier Foundation, writes: "There are myriad reasons why individuals may wish to use a name other than the one they were born with. They may be concerned about threats to their lives or livelihoods, or they may risk political or economic retribution. They may wish to prevent discrimination or they may use a name that’s easier to pronounce or spell in a given culture." https://www.eff.org/deeplinks/2011/07/case-pseudonyms

Pseudonymous speech has played a critical role throughout history, says York. "From the literary efforts of George Eliot and Mark Twain to the explicitly political advocacy of Publius in the Federalist Papers or Junius' letters to the Public Advertiser in 18th century London, people have contributed strongly to public debate under pseudonyms and continue to do so to this day."

A pseudonym can be name-shaped (e.g., "Jane Doe") or not. At time of writing, some websites - including Facebook - require that users use their "authentic identity," which typically means using your legal name or the name by which you are commonly known. This policy has caused some users, such as a group of drag performers in California, to lose their Facebook accounts. If we choose to use a pseudonym on social networks, it is important to understand that a risk of doing so is being reported for using a "fake name" and having one's account deleted. A strategy for avoiding that is using a name-shaped pseudonym. https://www.eff.org/deeplinks/2014/09/facebooks-real-name-policy-can-cause-real-world-harm-lgbtq-community

Persistent pseudonymity also offers us visibility. Visibility allows us to network with others, and by pinning our voice to a particular name we can develop an online reputation. An online reputation allows others to decide whether we are worthy of trust, and is therefore a crucial aspect in trust-based online communities. Reputation can be developed by consistently using a nickname or pseudonym that can either be connected to our legal identity, or not. The choice to connect our online reputation to our "real" name should be taken individually, according to needs and context. 

It is also possible to maintain multiple pseudonyms (and reputations) for different purposes. For example, a person involved in the gaming community and LGBTQI rights activism may wish to maintain separate identities for each purpose, and can build trust within each community separately doing so.

Collective Identity

Another way to be anonymous is through collective participation. This could mean a number of things, from a private group or mailing list that puts out collective statements, to a shared Twitter account. While the same security concerns apply, working from behind a collective identity means having the power of the crowd behind you, and can be a good option if you don't wish to reveal your identity. 

Comparing strategies

Whatever choice we make, what is important is that we keep our domains effectively separated.

No matter how many domains we identify in our digital life, and how many corresponding identities we create, on the internet every identity - even the one bearing our real or legal name - becomes a “virtual” persona and should be managed carefully. 

The pros and cons of the various identity options:

Real name

• Risk: Using your "real world" identity online means that you are easily identifiable by family members, colleagues, and others, and your activities can be linked back to your identity.

• Reputation: Others can easily identify you; gaining reputation and trust is easier.

• Effort: requires little effort.

Total anonymity 

• Risk: can be beneficial at times, but can also be very difficult to maintain. Choose this option carefully.

• Reputation: few opportunities to gain trust and reputation, or to network with others.

• Effort: difficult; requires caution. Might also require the use of anonymisation tools (for example Tor.)

Persistent pseudonymity

• Risk: Pseudonym could be linked to our real world identity.

• Reputation: A persistent pseudonym that others can use to identify us across platforms is a good way to gain reputation and trust.

• Effort: Maintenance requires some effort, particularly if we are also using our real name elsewhere.

Collective Identity

• Risk: possible exposure of our real world identity.

• Reputation: While not a way to gain individual reputation, you can still benefit from the reputation of the collective.

• Effort: Although secure communications are still important, requires less effort than total anonymity.

4. Creating a new online identity

What’s in a name?

The practice of naming varies greatly from one culture to another. While names, in one form or another, have existed across cultures for milennia, the concept of a "legal name" is a fairly new one.

Different countries regulate the practice of naming in different ways; for example, in Morocco, names must be chosen from a government-approved list, while in Germany a name must be clearly reflective of the baby's gender.

On the internet, platforms that have "real name" policies tend to base this judgement on an individual's legal name, rather than allowing them to identify as they choose. This can be problematic, not only for individuals trying to remain anonymous, but also for transgender individuals, individuals with mononyms, and others.

Because of such restrictions, it can be beneficial to select a "name-shaped" name when choosing a pseudonym. If we want to use commercial social networks, it is better to use a credible name and surname rather than more imaginative ones. Many companies will require that we use both a first name and surname, or a name that doesn't contain any slang terms or profanities.

Once we have decided on a name, a surname, and a username for our virtual persona, we should do thorough research - perhaps also using doxing tools and techniques (see the section on Self-Doxing) - to find out if someone else is already using that name. After all, if we wish to develop our own reputation, we don’t want to be confused with someone else, especially if they don’t share our views of the world!

Writing our own story

The practice of “story telling” (and of creating a social mask, for that matter) is an old one, and creating a new persona with a story makes it a lot easier to maintain the role.

We can base our story on a “known” person’s story, a superhero, a fictional character from our favourite novel, or adopt a “group identity” like Anonymous/Anonymiss or the Guerrilla Girls. If we feel particularly inspired, we can invent a new story.

The main point is that when we create an identity we should conceive a whole virtual persona, an avatar that needs to be nurtured and developed in order to become credible.

This page offers some helpful tips for inventing a new identity: Roleplay

Creating a credible persona

A virtual persona or identity can't be just a name with a mail address and a series of web accounts. If we keep all our normal identifying traits - such as our gender, job, attitude or the way we write - it might be possible for someone to connect the dots and connect our pseudonymous personas with our real identity.

• Linguistic fingerprint: One of the things to consider is our linguistic fingerprint. This could be identified through a so-called "stylometric analysis" using tools that are increasingly usable, even for non-experts, that make it possible to identify the author of a particular text. We can, for example, give away our real identity through our particular way of writing certain words, our typical typos, and our style and tone. To change this, we can start by using a spell-checker in our word processor to check for consistent typos. We could also think about adopting a different writing attitude. To keep it simple, we could adopt one simple rule for each persona, e.g. making them shout by only using capital letters, or be a low-talker with a lower-case style, or very excitable, with a lot of exclamation marks, or a spelling criminal, always putting apostrophes in the wrong place or mistyping words. What we decide to do with our writing style should match the character we choose for our new identity.

• Work: To be sure that our pseudonymous avatar cannot be tracked back to us, our persona should have a job that is different from ours, but not so different that we don’t know anything about that field: for example, they shouldn’t be a surgeon if we don’t know anything about anatomy!

• Skills and interests: Similar considerations should be made to select our persona's skills and the main topics they focus on and write about. These can, of course, overlap - for example a journalist who writes about national politics may have an alternative identity that talks about national politics as well, but only casually (and their main interest will be something different, like carpentry or feminism or food).

• Psychological attitude: As for psychological attitude, a good rule of thumb is to give our persona depth by creating some "weak spots" - but choosing them carefully so that, if the weak spot is attacked, we are able to weather the strikes and even have some fun in the process. For example, if we have a very strong sense of humour, we could choose a severe lack of humour as our persona's biggest weakness, so that if that weakness comes under attack we can enjoy impersonating a humourless persona without really being psychologically harmed.

What is most important is to remember that on the internet, each one of our identities - even the one connected to our real name - is a “virtual” identity, and it is always better to decide what character traits we want to expose in each of them. Creating a somewhat fictional character can be a good idea even for our “real” online identity.

More about how to create a rounded character for our identities here: https://lilithlela.cyberguerrilla.org/?page_id=94049

Creating online profiles

Once we have created several personas, it's important keep them separate in both our physical and digital lives. While keeping notes on our identities might help ensure that we remember our story, there are technical measures we can take to make sure that our profiles stay separate.

A good start is to create different browser profiles, mailboxes and social network accounts for each of our identities. A good rule of thumb is to always use different apps for each account/identity and, if possible, to separate our identities per device or operating system (see 6. A different machine for each identity).

Creating separate browser profiles and mailboxes

Browser profiles: To create multiple profiles with Firefox, visit: https://developer.mozilla.org/en-US/docs/Mozilla/Multiple_Firefox_Profiles . For Google Chrome, visit: https://support.google.com/chrome/answer/2364824

Mailboxes: When creating a new mailbox, it is always a good idea to connect to the server’s website with Torbrowser and, if a contact email address is required, to think about using a disposable email address instead.

Disposable email addresses

For some activities and social domains we need to manage rounded personas, in order to gain a strong reputation and trust from other members of the community. In some cases, however, all we need is a disposable email address that we only need to use once or few times, for example for opening an account in an untrusted platform.

Even if we decide to have just one identity online, using disposable email addresses prevents sites from building up a history of our activities and ensures that if that account gets compromised we can simply delete it and create a new one, keeping our digital life intact.

• Creating a disposable email address: There are many services that offer disposable email addresses. Some of the most privacy-oriented are: https://anonbox.net, offered by the Chaos Computer Club, and https://www.guerrillamail.com

• Tools to generate personal details: Making the process of creating a disposable account easier, Fakena.me (https://fakena.me) is a privacy-oriented '"fake name generator" that provides everything for you - from a credible name, birth date and (US-based) address, to a user name and password and a link to the connected guerrillamail mailbox. Another similar service, called Instant Internet Decoy (https://decoys.me) creates convincing but entirely fictional people who have birthdays, locations in several countries, families and even answers to common security questions.

• Using existing disposable email addresses: We can also make use of existing fake or disposable email accounts. BugMeNot(http://bugmenot.com) allows people to share their email logins and passwords created for platforms with free registration, for anyone to use.

Mail aliases

Another option is to create a mail alias - a different email address that is connected to our main mailbox. The advantages of this approach are that this email account will not expire, and if it gets compromised we can just dispose of it and create a new one. But of course if the alias receives a lot of spam, it will fill our main mailbox.

While not every mail service allows users to create mail aliases, this service is offered to every mail user of Riseup (https://we.riseup.net) and Autistici/Inventati (https://www.autistici.org), two secure, autonomous servers that are particularly focused on the right to privacy and anonymity.

Managing our identities on social networks

Note: When we use social networking platforms, we should always access them with a secure HTTPS connection.

• When creating an account for a new persona on a social networking platform, use the browser profile you have created for that persona. Make sure to check the privacy settings so that you know what you are making public, who can see what you post, who can contact you, who can look you up and what your contacts can do (can they tag you in pictures? can they write on your "wall"?)

• Also be very careful about the profile information you provide, as well as the profile picture and cover photo you use, as these are generally publicly available to anyone who looks for us in that social network, regardless of our privacy settings.

• Make sure your contacts do not overlap with your other identities, and to make sure your different identities don't "follow" one another. It is particularly not a good idea to follow your pseudonymous personas with your real identity. If someone is looking to unmask one of these personas, the first thing they will look for is who the account follows, and who follows the account. For the same reason, we should avoid reposting posts or other content published by one account with another account.

• Most social networking platforms will display your location where they can. This function is generally provided when we interact with the platform using a GPS-enabled phone, but it can also happen if we are connecting from other devices - for example the network our computer is connected to may also provide location data. It's always a good idea to double-check your settings - particularly on photo and video sharing sites.

• Photos and videos can also reveal a lot of information without you realising it. Many cameras will embed metadata into your photos, which can include the date, time and location of the photo, camera type, etc. Photo and video sharing sites may publish this information when we upload content to their sites.

• If we access social networking platforms via mobile apps, it is better to use a different app for each separate account, so as not to post something to the wrong account by mistake. There are several apps which can be used to manage your social networking platforms - it is, however, a good idea to use a different one for each identity, to reduce the risk of giving away your real identity.

• Publish from your various identities at different times of the day. Some social networking platforms, like Facebook, allow users to schedule the publication time of their posts. To learn how to do this, read:https://www.facebook.com/help/389849807718635 . Buffer (https://buffer.com) and Postcron (https://postcron.com) enable you to schedule a post on Twitter and other platforms.

• It can be a good idea to follow, from our pseudonymous profiles, other people who might reasonably be considered the real owners of that profile. To further distance our real identity from our pseudonymous identities, we can also write (and hashtag on Twitter) posts under our pseudonymous profiles about events that we are not attending, especially if they are taking place far away from us. It can also be fun to publish and then delete posts that look like we have exposed our identity, so as to further confuse anyone who may try to unmask us.

• Whatever social networking platform we decide to use, you should always read its terms of service to check if they suit your purposes. To get a summary of the terms of service of many social networking platforms (and other web services), go to the website Terms of Service; Didn’t Read (https://tosdr.org).

Alternative social networks

Mainstream commercial social networking platforms like Facebook or Twitter can be extremely useful if our aim is to publicise as widely as possible an event we are organizing or a project we are launching.

However, if you're using one of these platforms it is important to be aware that:

• these platforms have very strict terms of service that could justify their decision to close our accounts if they find that our contents go against their rules (for more, read: http://www.aljazeera.com/indepth/opinion/2013/05/20135175216204375.html)
• users of these platforms are profiled, and information is sold to advertisers. If we add to this the ever-changing terms of service and the interactions with other apps and features that make it very difficult to understand clearly what actually happens to our data, the best solution is to limit the use of commercial social networking platforms to specific projects we want to publicise to a wide audience.

There are also alternatives available - social networking platforms that give much more freedom to their users and don't profile them.

• Social networking platform Ello explicitly states in its manifesto that “You are not a product”. It also does not require "real names" - a rarity amongst commercial social networks. It's important to be aware, however, that Ello is still a commercial project
• There are alternatives that are community-based, distributed rather than centralized, based on free and open-source software and privacy-friendly. Among these, Diaspora (https://joindiaspora.com), Friendica (https://friendi.ca) and Crabgrass (https://we.riseup.net) are especially worth mentioning.
• Other similar sites may be popular in different regions, so we way wish to explore other options.

Before choosing to use a social networking platform, we should ask:

• Does it provide connection over SSL (like HTTPS) for all uses of the site, rather than just during login? Are there any problems related to encryption (eg related to encryption certificates)?
• According to the platform's End User Licence Agreement, Privacy Policy and/or Data Use Policy, How is your content and personal data treated? With whom are they shared?
• What privacy options are provided for users? For example, can we choose to share our videos securely with a small number of individuals, or are they all public by default?
• Is the geographical location of the servers known? Under which territorial jurisdiction do they fall? Where is the company registered? How does this information relate to the privacy and security of our activity and information?

Creating a blog or website

Writing from our social networking accounts - especially if we have found a more privacy-friendly option - may be all we need to voice our opinions and to interact with a given network of people (or with several networks, if we are using different identities online). On the other hand, we may want a more independent platform to spread our ideas, plans or creations. For this we can create a website or a blog.

• Creating a blog can be as easy as signing up to a blogging platform and choosing a name and a "theme" or visual template. There are several blogging platforms that are both user-friendly and free, including the open-source Wordpress (Wordpress.com).

• Based on Wordpress, but with some tweaks for additional user privacy, are two security-oriented platforms that are managed by autonomous servers: Autistici/Inventati’s (A/I) Noblogs (http://noblogs.org/ http://noblogs.org) and Nadir’s BlackBlogs – (http://blackblogs.org/ http://blackblogs.org). To create a blog on either of these platforms, all that is needed is to have an email account hosted on an autonomous server (see https://www.autistici.org/en/links.html and https://blackblogs.org/policy respectively for a complete list).

• If we want a complex graphic layout or need to install particular tools that are not offered by Wordpress and its plugins, we can create our own website. For this we need to get some space in a server through a webhosting service. There are many webhosting services out there, but since they generally aren’t free, the options to stay completely anonymous are reduced to creating a website with A/I, which by default does not connect the users of its services with real identities. To learn more about Autistici/Inventati’s webhosting service, visit: https://www.autistici.org/en/services/website.html

• If we want to use our own domain name, bypassing payments and identifications may get difficult unless we use Bitcoin or another anonymous payment system. The personal data we will provide will not only be stored in the registrar’s internal archives, but by default will also be recorded in a database that can be easily queried by anybody through a simple command in a search engine (whois) or on several websites such as Gandi.net (https://www.gandi.net/whois). To avoid this, we can register our domain with the data of an association and use a prepaid credit card that is not connected to our own data (if available in our country). Alternatively, we can use a registrar like Gandi.net (https://www.gandi.net) that offers private domain registration for individuals whenever possible.

5. Managing collective online identities

Collective virtual personas

General Ludd, Captain Swing, the Guerrilla Girls, Luther Blissett, Anonymous - for centuries groups and like-minded people have participated anonymously in historic protest movements, or have created ground-breaking and provoking artworks or pranks under a collective pseudonym. Besides hiding the identities of the individuals involved, these collective personas have shrouded their feats in an aura of myth and almost magical power.

If we want to adopt a collective identity, we can adopt one that already exists, like Anonymous/Anonymiss, or create a new one that we can then share.

• A very interesting study on collective identity is anthropologist Gabriella Coleman’s book Hacker, Hoaxer, Whistleblower, Spy. The Many Faces of Anonymous (2014), based on an anthropological research on how Anonymous became a well-known and powerful collective identity with multiple faces and attitudes.

• In her PhD dissertation Networked Disruption. Rethinking Oppositions in Art, Hacktivism and the Business of Social Networking (2011), Tatiana Bazzichelli describes how multiple identities have been used to disrupt the fundamental notions of power and hegemony on which Western culture is based, and how this works in the web today.

Managing collective identities... or simple collective accounts

While the collaboration of many individuals can help create a rich collective identity, managing a collective project may have some security and technological challenges we should keep in mind.

Since a single weak link in a security chain can break the entire system, our security and anonymity depend on the precautions each member of our group takes. We may decide that we don’t want to be absolutely anonymous; that our close friends can know about our collective activities. The degree of security we may want to attain for our group depends on the possible threats we face and on our adversaries’ power and skills.

Nonetheless, there are some important things we don't want want to lose, like the password to our collective mailbox or to our group’s social networking accounts. If we decide that we are going to share those passwords with the whole group, each member needs to be trained on how to store a password securely.

To minimise this kind of risk, we should try to use services that provide for different accounts and passwords whenever possible. For example, instead of using a single mailbox, we may create a mailing list that all the group members subscribe to. If we allow non-subscribed people to write to it, each group member will be able to read that e-mail in their own mailbox rather than in a collective one with a dangerously shared password.

Similarly, if coordination really needs to happen through Facebook, it is much better to share information in a dedicated group rather than do everything within a collective account.

6. A different machine for each identity

If we use the same operating system for our several identities, no matter how carefully we separate our profiles we can still make a human mistake by, for example, connecting to a pseudonymous account through the browser profile we have assigned to our true identity, or get infected by a malware that allows our attacker to monitor everything we do online, with all our identities.

Both risks can be limited by using a virtual machine for each of our domains, and by reserving yet another virtual machine to opening untrusted attachments in order to avoid a malware infection.

As the name suggests, a virtual machine (VM) is basically a simulated computer with its own operating system, which runs as software on our physical computer. We can think of a VM as a computer within a computer. Installing and running a virtual machine is not very complicated, and there is very good documentation around. For our purposes of anonymisation, the best available option is to install Virtualbox, an open-source, cross-platform virtual machine monitor (https://www.virtualbox.org). Using Virtualbox, you can create a virtual machine, and then run an operating system on it called Whonix.

Whonix: compartmentalizing our identities through a secure virtual machine

Whonix (https://www.whonix.org) is an operating system that aims at protecting our anonymity, privacy and security by helping you to use your applications anonymously. A web browser, IRC client, word processor and more come pre-configured with security in mind. Whonix is a complete operating system designed to be used in a virtual machine. It is also free software, based on Tor, Debian GNU/Linux and security by isolation.

Whonix’s website offers a wide documentation, ranging from very clear installation and usage instructions to thorough recommendations on security and the risks we may run: https://www.whonix.org/wiki/Documentation

Tails: a live system that leaves no traces

Using virtual machines, Whonix in particular, is a good idea if we have our own computer where we are free to install whatever we want, especially if it is a powerful machine. But if we use an older box or just connect to the internet from cybercafes, installing a virtual machine becomes unsustainable. In this case, we may turn to a live Linux distribution, a USB-stick with a Linux operating system installed on it that runs in the computer we are using as soon as we switch it on.

A particularly secure live distribution focused on security and anonymity is Tails, or The Amnesic Incognito Live System, a free and open-source Debian-based Linux distribution that can be started on almost any computer from a DVD, USB stick, or SD card and forces all its outgoing connections to go through Tor, blocking direct, non-anonymous connections.

When we launch Tails, we have a complete operating system that, just as Whonix, comes with several built-in applications pre-configured with security in mind: web browser, instant messaging client, email client, office suite, image and sound editor, etc. With Tails, we can access the internet, communicate, and do all we need anonymously and securely and, after the computer is shut down, the system will leave no traces on the machine unless we ask it to do so.

When using different identities, it is not advisable to use the same Tails session to perform two tasks or endorse two contextual identities that we really want to keep separate from another, for example hiding our location to check our personal email and publishing a document on our pseudonymous blog.

The first reason is that Tor tends to reuse the same circuits, for example amongst a same browsing session, making it easier for a powerful adversary to correlate the several browsing requests as part of a same circuit and possibly made by a same user. Second, in case of a security hole or a misuse in using Tails or one of its applications, information about our session could be leaked. That could reveal that the same person was behind the various actions made during the session.

The solution to both threats is to shut down and restart Tails every time we are using a new identity.

Security by isolation: Qubes OS

There are three approaches to digital security: the first one is security by obscurity, which is based on encryption, strong passwords and similar measures and acts as a first line of defence, as a deterrent that will discourage random attacks but is not likely to stop someone who is directly targeting us; then there is security by correctness, whereby software developers try to get rid of bugs that make their code vulnerable. But modern software is very complex, and it is almost impossible to do this job perfectly. Therefore, the most pessimistic, and realistic, approach is security by isolation, which gives for granted that security measures can be vulnerable and focuses on harm reduction by stopping possible attackers from accessing the whole system that we want to secure.

In order to implement security by isolation, for instance, Whonix is divided into two parts: Whonix-Workstation, which is the system we access for our work, and Whonix-Gateway, which routes all internet traffic through Tor and, by being isolated from the workstation, averts many threats posed by malware, misbehaving applications, and user error.

But while Whonix needs a virtual machine to run on and its main focus is anonymization, there is an operating system that has security by isolation as its main purpose and that can make life a lot easier to someone who manages several social domains and/or identities in her digital life.

Called Qubes OS – https://www.qubes-os.org – and developed by computer security researcher Joanna Rutkowska, Qubes is a free and open-source security-oriented operating system based on Fedora, a Linux distribution, and Xen, a virtual machine monitor (or hypervisor), that allows us to separate the various parts of our digital life into securely isolated virtual machines. Qubes keeps the things we do on our computer securely isolated in these different VMs so that if one virtual machine gets compromised, the other won’t be affected. This way, we can do everything on a single physical computer without having to worry that one successful cyberattack harms our whole system, potentially revealing all the connections among our several identities.

A plus of Qubes OS for the purposes of multiple identities management is its user-friendly window manager, that assigns a different colour to each domain. Thus, the colour of the frame makes each window clearly recognizable as belonging to the domain corresponding to that colour and prevents potential human mistakes in the management of our identities.

Tails, Whonix, Qubes OS: how to choose

The three tools we have described in the last few paragraphs – Whonix, Tails and Qubes OS – all allow us to use a completely separate operating system for managing our alternate identities, and can be quite useful to make sure that we don’t reveal our true identity while we use the anonymous one(s). Still better, if we can afford it, would be to have a different machine for each of our identities, as well as an air-gapped one to store our most sensitive data. Of course, the choice depends on the resources we can dedicate to securing our digital life and on our threat-model, especially on the adversaries we expect to face: if we are pretty sure that in our country harassers work together with the regime to slander (or worse) people like us, we may want to be absolutely sure that none of our data is leaked and – if we have enough funds or are connected to a network of hacktivists – we may decide to ask for an expert’s help to check that we have taken all the necessary measures to keep doing an efficient work and to stop any ill-intentioned actors from tampering with our data. On the other hand, if we feel sufficiently protected by our community and/or by our government, and we expect our adversaries not to be very skilled in technological matters, perhaps all we need is to separate our browser profiles and mailboxes, to use a secure HTTPS connection and some basic common sense and to keep developing a network of support for vocal women online.

If we think that using a different operating system is really necessary, or that it can be helpful to keep things logically and graphically separated (or we just want to explore all the possibilities we have for the fun of it!), we should assess our resources and our needs in order to identify the best solution for us.

Comparison between Tails, Whonix and Qubes Os:

As we can see in the table above, Tails just needs a normal computer and a DVD or a bootable device to launch the system, but installing the system in this device, as well as having the DVD, USB stick or SD card actually boot in the computer we are using, can be tricky, and we may need some external help. After that, though, using Tails is pretty easy, and if what we need to do with our alternate persona needs a focus on anonymization, then it may be worthwhile to overcome the initial obstacle. Tails is a good option also if we have few resources, if we don’t have a computer of our own, or if we often use computers at internet cafes and want to be safer. One particular advantage of Tails is that after we have switched the computer off, we leave no traces and everything we have done vanishes into thin air. But if on the other hand we need persistence, i.e. we want to keep some files we have created or downloaded in our USB stick or we have changed some system settings and want to keep them also in the future, we need to enable this option when we start the system. Last but not least, Tails is an established, respected project that has been developed for many years and is used by a wide community of people.

If what we need is both anonymity and security by isolation and we have a good machine – https://www.whonix.org/wiki/System_Requirements – where we can run Virtualbox, installing Whonix in one or more virtual machines, according to the number of our alternative identities, seems a good solution that caters to all our needs and also offers an excellent documentation: https://www.whonix.org/wiki/Documentation. Nevertheless, Whonix, like Qubes Os, is a relatively recent project and the community using it is still rather small.

Qubes OS is a good choice if we want to keep all our activities inside our own computer without having to install anything else and if what we are trying to do is to effectively separate our identities rather than anonymize our activities in the web. It requires a very powerful computer – http://qubes-os.org/trac/wiki/SystemRequirements – and this can be a hindrance, but if we feel that we really need to protect ourselves against possible cyberattacks, the investment may be worth its while.

To sum up, none of these tools protects us from every threat, and we shouldn’t look at them as a magic potion that will make us invulnerable. Nevertheless, by using any of them, according to our needs and resources, we will raise the level of effort that an attacker will need to harm us, thus making an attack less likely.

For a wider comparison among these and other systems, go to: https://www.whonix.org/wiki/Comparison_with_Others