The Facebook whistleblower's testimony was refreshing and frightening by turns, revealing the company's awful internal culture, where product design decisions that benefited its users were sidelined if they were bad for its shareholders.
The big question now is, what do we do about it? The whistleblower, Frances Haugen rejected the idea that Facebook's power should be diminished; rather, she argued that it should be harnessed - put under the supervision of a new digital regulator.
Regulating tech is a great idea (assuming the regulations are thoughtful and productive, of course), but even if we can agree on what rules tech should follow, there's still a huge debate over how the tech sector should be structured.
Like, should we leave monopolies intact so that we only have to keep track of a few companies to make sure they're following the rules? Or should we smash them up - through breakups, unwinding anticompetitive mergers, and scrutinizing future mergers?
For me, the answer is self-evident: if we don't make Big Tech weak, we'll never bring them to heel. Giant companies can extract "monopoly rents" - huge profits - and cartels can agree on how to spend those profits to subvert regulation.
We need to fix the internet, not the tech giants.
The problem isn't just that Zuck is really bad at being the unelected pope-emperor of the digital lives of 3,000,000,000 people - it's that the job of "pope-emperor of 3,000,000,000 people" should be abolished.
I believe that people who rely on digital tools should have the final say in how those tools serve them. That's the proposition at the core of the "nothing about us without us" movement for accessible tech, and the ethos of Free Software.
Technologists should take reasonables step to make their products suitable for users, and regulators should step in to ban certain design choices: for example, algorithms that result in racial discrimination in housing, finance and beyond.
The law should step in when sites or apps are deceptive or fraudulent or otherwise harmful; people hurt by negligent security and other choices should have remedies in law, both as private individuals and through their law enforcement officials.
But even if we did all that - and to be clear, we don't - it wouldn't be enough to deliver technological self-determination, the right to decide how the technology you use works.
For example, when the W3C was standardizing EME - a shameful incident in which they created a standard for video DRM - there was a lot of work put into accessibility, including safeguarding closed captions and audio description tracks.
But even the most inclusive design process can't contemplate all of the ways in which users will need to adapt their tools. My friend Jennifer has photosensitive epilepsy and was hospitalized after a strobe effect in a Netflix stream triggered a series of grand mal seizures.
EME could have accommodated that use-case by implementing a lookahead algorithm that checked for upcoming strobes and skipped past them or altered their gamma curves so that they didn't flash.
The committee rejected this proposal, though.
But that wasn't all they did. They also rejected a proposal to extract a promise from the companies involved in EME's creation to refrain from threatening toolsmiths who added this feature on their own, either to help themselves or on behalf of other users.
The companies making W3C DRM said they didn't need to promise not to destroy the lives of toolsmiths who added accessibility features to their product because they would add every necessary accessibility feature themselves.
Except they wouldn't. They blocked an anti-seizure tool, and Dan Kaminski's proposal to shift color palettes to compensate for color-blindness, and a proposal for captioning tools to bypass DRM to ingest videos and run multiple rounds of text-to-speech analysis.
Even if they *had* accepted all of this, it wouldn't have been enough. No one can anticipate all the ways that people need to adapt their tools. "Nothing about us without us" can't just mean, "Some people with disabilities helped design this."
It also has to mean, "I, a person using this tool, get a veto over how it works. When my interests conflict with the manufacturer's choices, I win. It's my tool. Nothing about me without me." That's the soul of technological self-determination.
In response to the whistleblower's testimony, Slate Future Tense has an article by Louis Barclay describing his misadventures with a Facebook add-on he created called "Unfollow Everything" - a tool that helped people use Facebook less and enjoy it more.
Unfollow Everything unfollowed all the friends, groups and pages you followed on Facebook. This eliminated Facebook's News Feed entirely, and then users could either manually check in on friends (unfollowing isn't the same as unfriending) or selectively follow them again.
Users who tried Unfollow Everything really liked it.
It would be interesting to understand this phenomenon. After all, Facebook has spent a lot of money on internal experiments where social scientists and designers collaborated to increase the time users spend on Facebook (AKA "engagement").
In his spittle-flecked rebuttal to the WSJ's Facebook Files, Mark Zuckerberg insisted that the point of this research was to increase user satisfaction, and the extra ad revenue from additional pageviews were just a coincidence.
That study never happened and never will, because Facebook had Unfollow Everything deleted from the Chrome store, kicked Barclay off the service and threatened to rain legal hell upon Barclay's head if he ever made *any* Facebook tools, *ever* again.
Like the companies that standardized DRM at the W3C, Facebook says its users' wellbeing and satisfaction are primary factors in its product design and demands a veto over any modifications so it can deliver the best product to those users.
Like those companies, Facebook opposes - and wields its legal might against - toolsmiths who adapt its products and services to reflect the needs of users when those needs conflict with its shareholders' interests.
Time and again, Facebook demonstrates that it cannot be trusted to wield a veto over how we use Facebook. Think of how the company is attacking Ad Obervatory, an academic project that recruits volunteer Facebook users to track paid political disinformation on the platform.
Facebook falsely claimed that Ad Observatory compromised users' privacy, and directed researchers to its own tools for analyzing content on the platform - tools that are riddled with errors and omissions and flat-out misleading distortions:
(To be fair, Facebook *did* once have a reliable research tool, Crowdtangle, which produced accurate and thus unflattering data about FB, so Facebook neutered the tool, disbanded the team that worked on it, and forced the founder out of the company)
As part of its smear campaign against Ad Observatory, Facebook argued that it *has* to shut down this kind of activity, because that's how it defends us from privacy-invading scumbags like Cambridge Analytica.
If the specter of platform abuse - racial discrimination, harassment, infringement, malware, fraud, privacy invasion - grants Facebook a veto over independent modifications of its products, FB will fail to protect us, *and* will kill the tools that *can* protect us (from Facebook - and other bad actors).
And yet, FB really is used in ways that harm its users. Sometimes, the abuse comes from third parties that create plugins, bots, scrapers or other interoperable tools that modify Facebook. But you know what? Many of these abuses come from Facebook itself.
That's why Facebook can't be in charge of the Facebook Cops. That's why FB can't have the final say over how we modify Facebook (and why no tech company should have the final say over how its users adapt the things they buy and use).
Which brings me back to the whistleblower and her call for regulating Facebook. Yeah, we need to do that. We need to have things like a federal privacy law with a private right of action, robust protections against harassment, vigorous enforcement of anti-fraud rules.
That's how you distinguish between good mods and bad mods - by passing democratically accountable laws that tell us what we are and are not entitled to.
Not by letting Facebook (or Apple, or Google, or Netflix) decide what is and isn't acceptable.
This is the thesis of "Privacy Without Monopoly," the EFF white paper Bennett Cyphers and I wrote on how to balance interoperability and privacy: we *do* need privacy protections, *and* we can't rely on monopolists to decide what those should be.
For example, a settlement with FB could require it to seek permission from a "special master" before it could threaten to sue an interoperator like Ad Observatory, Unfollow Everything, Friendly Browser, GBWhatsApp, or any of the many other independent tools FB has pursued.
The special master could determine whether FB was engaged in privacywashing (or securitywashing, or safetywashing, etc) in a bid to make harming its users to help its shareholders seem legitimate.
Users don't always make wise choices about how to use their tools, but tech giants can't be trusted to distinguish between "preventing harms to users" and "preventing harms to the bottom line."
That should be obvious, but if there's any doubt, surely the Unfollow Everything scandal settles the matter.
@pluralistic A few lessons that are forgotten in the Cambridge Analytica saga: the UK's ICO found that the whistleblowers were full-of-it, and that Cambridge Analytica's "technology" amounted to script kiddies with scikit-learn.
I happen to disagree somewhat --- even weakly-effective tools of persuasion can be significant and determinative where margins of difference are thin and consequences of a few votes or voices large.
But both objections still leave us with that inconvenient fact that manipulation is what CA were selling, and what its customers were buying. Moreover there's the age-old well-tested value of Kompromat and blackmail material enabled through massive surveillance, or even possibly just the plausible threat that such material exists. Barbed wire and prods don't have to kill to lead cattle to slaughter, just inflict enough pain.
CA swiped gobs of personal information from Facebook, with Facebook's full support at the time, and turned around and attempted to sell it to the least principled of actors for underhanded and deceptive purposes. If CA happened to be cheating its own customers at the same time it was cheating Facebook members and public elecorates ... well, honour among scoundrels and all that.
Mamot.fr est une serveur Mastodon francophone, géré par La Quadrature du Net.