Notes from the Internet Health Report 2019

This week I read (the print-version of) the Internet Health Report 2019, published by the Mozilla Foundation about issues that potentially stand in the way of a web that ‘puts people first’. With the phrase ‘putting people first’, the report means ensuring aspects like safety, openness, inclusion and control. Some of these things improved in the past years, some less so than one would have hoped.

I read it by the waterside on my holiday

The Internet Health Report is edited by Solana Larsen for Mozilla Foundation and puts together input from hundreds of readers and experts in an extremely readable format. Or multiple formats, really, as there is the website, the ebook and (new this year) a printed edition. The print version definitely helped me read the full thing. Everyone is invited to participate:

this publication is neither a country-level index nor a doomsday clock. We invite you to join us in assessing what it means for the internet to be healthy, and to participate in setting an agenda for how we can work together to create an internet that truly puts people first

(From the README)

In his update, Mozilla Foundation’s Mark Surman calls out three things that improved: calls for privacy are becoming more mainstream, people start to speak up about and work on more responsible AI and a healthy discussion about big tech is taking place among more and more people. Some things also got worse, Surman writes: there is more internet censorship, biometrics are being abused and increased AI usage specifically affects minorities because of bias, despite efforts like ethics boards.

This post has some of the things that piqued my interest, with links. For the full content, including all sources and links, get the printed edition or visit the Mozilla Internet Health Report 2019 website, where you can read everything or just the things you’re interested in. There is also a PDF of the short version.

Data: this year’s spotlight

In the spotlight section, the Internet Health Report features three articles that all have something to do with data.

AI

‘Use of AI is skyrocketing (for fun, as well as for governance, military and business) and not nearly enough attention is paid to the associated risks’, says the report. Governments start to use AI in military, for immigration decisions and for surveilling citizens, places one can imagine (racial) bias can cause serious harm. OpenAI decided not to release a trained data set as they worried about malicious uses of the technology. Some companies founded ethics boards to govern their AI usage, but they are often big companies with, inevitably, competing priorities. See also: Let’s ask more of AI

Cities

Data is useful for cities to base their policies on, and are somewhat envious of the ‘data monopoly’ big tech companies like Uber and AirBnB have. Some only give permission to do business on the condition of getting access to businesses’ data. See also: The power of cities

Ads

Paying for products with your personal data threatens freedom and human rights, because the algorithms that let companies do targeted advertising, can also be exploited by people with bad intentions. Some change is happening: Facebook, Twitter and Google all took some action against abuse after public pressure and new legislations. Lots of browsers now protect against tracking and/or block ads. Companies also start to realise targeting and knowing everything about everyone doesn’t necessarily make them more effective. See also: Rethinking digital ads

Safety

More and more awareness around privacy on the web is raised; by the campaigns of non profits like the EFF and Mozilla, but also by numerous data leaks that got into mainstream news and GDPR, the European data protection legislation that affects pretty much all businesses that are online. Tracking protection has become something browsers compete on (Firefox and Safari are ahead in this game), and even Facebook’s CEO announced that they now care about privacy.

• While anonymity online can be abused by criminals, it is also absolutely worth defending and an important tool for whistleblowers reporting on corruption, or for people whose government oppresses them. See also: In defense of anonymity
• If you don’t want big tech to have your data, you just don’t give it, right? Well, Katarzyna Szymielewicz has a useful metaphor of three layers of data: the first is data you share and control, the second is behavioral data and metadata that others collect (if you want to control this, you need expertise) and the third is what machines think about you (we have no control, because they are the result of companies combining data with algorithms we often can’t see). See also: Show me my data and I’ll tell you who I am

Openness

The web has always been fundamentally open. Imagine starting a new website required a government issued permit, or that getting onto any website required a certain kind of education, nationality or payment… Worldwide, this openness is threatened, as there exist walled gardens, social media taxes and shutdowns.

• Governments who want to frustrate their citizens’ communications used to turn off the whole internet, but this was relatively easy to detect (for instance, by organisations protecting human rights), so they try out other tactics, like shutting down the internet in just one region or slow it down instead of turning it off. See also: Internet slowdowns are the new shutdowns
• Various African countries now have taxes on using social media and messaging apps (Uganda), VOIP (Zambia) and even blogging and vlogging (Tanzania). See also: Taxing social media in Africa
• Germany implemented a law against online hate speech and harrassment, which can force social media companies to take down certain content. Some call for web companies to be open about who reported content and how complaints are handled. See also: Inside Germany’s crackdown on hate speech
• Wikidata, a project by the non-profit behind Wikipedia, offers lots of volunteer-created structured data. Technology firms who make voice assistants get a lot of value from this. Despite that, vast majority of donations is still from individuals, and only 4% from corporations. See also: Wikidata gives wings to open knowledge

Inclusion

Is the internet a level playing field yet for people from all backgrounds? This question has multiple sides to it: it is about how online communities fight harrassment and abuse, and about the working conditions of people who make internet hardware.

• Women and women of colour in journalism are (still) much more likely to be harrassed online than men, various study showed. See more: Women journalists feel the brunt of online harassment
• Contributors to open source projects that power a lot of the world are often a homegeneous group. That’s bad as code is not neutral and i.e. biased towards the people who write it. Strictly enforced code of conducts are appreciated by underrepresented groups, have helped community members call out bad behaviour and are therefore a helpful instrument in trying to make open source communities more diverse. See also: Codes of Conduct now guide open source communities

Web literacy

Do people understand the web well enough to make informed choices about things like sharing baby photos and recognising fake news? Do we understand it well enough to use it to people’s benefit, for example as a platform for activists to collaborate on?

• 3101 people from 124 countries worked together analysing footage, interviews and expert analysis in a project called Decoded, to show that the US-led coalition’s bombing of the Syrian city of Raqqa costed not 23 civilian deaths, but over 1500. See also: Decoding images of war in Syria
• User tracking and targeting employed for manipulation of public opinion with fake news threatens democracy. Cambridge Analytica was not only involved with influencing the British and US electorates, but also worked in many other countries, including Kenya, Brazil and Mexico. For this year’s European Election, Google, Twitter, Facebook and Mozilla pledged to do somethimg about this, signing the European Commission’s Code of Practice on Disinformation. See also: The challenge of democracy in the digital era
• Are we addicted to the internet? Research that showed Americans spend 6 hours a day on their devices and (other research) showing maintaining a user’s attention as a design principle seem to suggest so. That’s probably not time well spent. In the meantime Apple and Facebook introduced features to limit screen time. See also: Breaking free of the addiction machine

Control

Should the web be controlled by few or many? In other words, should power lie with a couple of companies that everything centers around, or do we want the web to be more decentralised? The Internet Health Report 2019 contains lots of good arguments for decentralisation.

• There’s the question of ‘who (literally) controls the internet’. The cables deep under the sea that make the web work used to be built by telecom cariers in the early 90s, but are now invested in by private companies like Google, Facebook, Amazon and Microsoft (who, in 2018, owned or leased over 50% of these cables between them). See also: The new investors in underwater sea cables
• For Uber, and many ‘unicorn’ companies like it, the structure of ‘everything (taxi rides, etc) goes through our company’, ie centralisation, is what gets their investors’ hope for returns up. Fine, that’s capitalism. But the model gets in the way of others providing alternatives or putting users in control. This is an issue when such companies have become utilities, says Nathan Schneider. These companies would probaby do things very differently if their users owned them, in which case they would only have to consider user interests. See also: What if Facebook were owned by its users?
• The largest five internet companies make money by selling user’s attention to advertisers (Google, Facebook and Baidu), selling devices (Microsoft and Apple), being a middle man (Amazon and Alibaba) and doing lots of things from payments to in-app purchases to ads (WeChat parent company Tencent). See also: How do the biggest internet companies make money?
• Nextcloud is an effort to create an open source cloud, but unlike with traditional open source software it is not cheaper (free) compared to alternatives, it costs more (because you need to pay for hosting somewhere). See also: An open source alternative for “the cloud”

I learned a lot from reading the Internet Health Report 2019, from great new initiatives to serious problems that I wasn’t aware of. As a reader of this blog, I invite you to dive in yourself, too!