An underwater photo taken looking up to a large circular school of fish. The image is distorted by digital artefacts and a green filter is applied to the image.

How to Think About Making AI and Social Media Pro-Human

In this photo provided by Ukraine's 65th Mechanized Brigade press service, a Ukrainian serviceman launches a drone towards Russian positions at the front line in the Zaporizhzhia region, Ukraine, Wednesday, Nov 11, 2026. (Andriy Andriyenko/Ukraine's 65th Mechanized Brigade via AP)

AI Is Killing Autonomously. Time for New Rules.

A photo of a forest path forking in two directions.

You’ve Changed. Shouldn’t Your Algorithm?

People walk past Nasdaq video monitors displaying news of a sharp decline in Nvidia shares following reports that CEO Jensen Huang has resigned after revelations of extramarital relationships, Times Square, New York, NY, November 12, 2026. (Photo by Anthony Behar/Sipa via AP Images)

Despite Huang Scandal, US-China AI Safety Dialogue Still Seems on Track

Flags of the European Union aside a European Commission building.

A woman is passing by the logo of the European Commission at the Berlaymont building in Brussels, Belgium, on Friday, 17 April 2026. (Photo by Marius Burgelman/Belga/Sipa USA)(Sipa via AP Images)

The EU Promises Digital Fairness, But Still Wants Users To Do the Work

A close-up photo of a tree branch with an insect crawling on a twig. The image is distorted by digital artifacts. 

Distorted Branch by Lone Thomasky & Bits&Bäume / 

What the ‘AI Torture Chamber’ Debate Got Wrong About the Research

A protester holds a No Data Center sign Monday, Oct. 12, 2026, in Mason, Mich. (AP Photo/Paul Sancya)

Memo to New Governors and Legislators on Effective Data Center Reform

Rep. Jim Jordan (R-Ohio) looks on during a media availability after he was elected to be House Republican Leader for the upcoming 120th Congress on Capitol Hill on Nov. 10, 2026. (Francis Chung/POLITICO via AP Images)

Will a New Congress Ease the Pressure on Europe’s Digital Regulations?

Campaigners march against data centers in Dublin, Ireland, on November 6, 2026. Photo by Paula Geraghty.

Ireland is Helping Shape a Data Center Boom That Could Benefit Big Tech

Members of the Garda Armed Support Unit outside the National Gallery in Dublin as more than 50 world leaders will visit Dublin for the European Political Community (EPC) summit on Thursday, followed by an informal meeting of the EU Council on Friday. The biggest security operation in the history of the Irish state is under way, as authorities in Dublin mount a ring of steel throughout the capital. Picture date: Wednesday November 11, 2026. PA Photo. Photo credit should read: Niall Carson/Press Association IRISH (Press Association via AP Images)

Europe’s AI Future and Security Challenges Take Center Stage at Dublin Summit

OpenAI CEO Sam Altman delivers the keynote address during the OpenAI developers conference at Fort Mason on September 29, 2026 in San Francisco, California. (Photo by Heather Diehl/Getty Images)

Friends of the Court Flock to the OpenAI Copyright Fight

Sam Altman watches over a family consuming his AI. The living room vignette is framed by power tool imagery. The words “Behaviour Power” are the top of the artwork, framing the intent of the piece.

Behaviour Power II by Bart Fish & Power Tools of AI / 

Humans Have Rights. Machines Do Not.

A blue coloured icon of a data server, surrounded with comic-book style words 'Gasp!', 'Boom!' and 'Pow!' in colours of yellow and red.

Don't Let AI Become the Internet's Gatekeeper

The United Nations headquarters in New York. Shutterstock

The United Nations headquarters in New York. 

A Defining Moment for the United Nations and the Future of AI

An illustration of Tigrayan university chemistry Professor Meareg Amare by Alina Pechenckina. Source: Amnesty International / CC by 4.0

An illustration of Tigrayan university chemistry Professor Meareg Amare by Alina Pechenckina. Source: 

Meta Returns to Kenya’s High Court Over Hate Speech in Ethiopia

Executive Vice-President for Tech Sovereignty, Security and Democracy of the European Commission Henna Virkkunen, and European Commissioner for Democracy, Justice, the Rule of Law and Consumer Protection Michael McGrath speak at a press conference on Wednesday. Source

Executive Vice-President for Tech Sovereignty, Security and Democracy of the European Commission, Henna Virkkunen, and European Commissioner for Democracy, Justice, the Rule of Law and Consumer Protection, Michael McGrath, speak at a press conference on Wednesday. 

Assessing the EU Digital Fairness Act and Its Promise of Consumer Protection

A Close Read of Big TikTok and Meta Settlements Raises Important Questions

A collage of circuit boards interwoven with aerial photographs of water treatment facilities, agricultural lands, industrial sites and mining machinery.

Brazil's AI Development Strategy Needs an Edge Layer

Anthropic's logo is displayed on a smartphone with a stock market chart in the background on October 27, 2026, as Anthropic filed a Form S-1 with the Securities and Exchange Commission (SEC) for a planned initial public offering in the United States. (Photo by Samuel Boivin/NurPhoto via AP)

Could the United States Seize Anthropic at a Discount?

Journalists test AI software used to manipulate images at Phnom Penh Municipal Police headquarters following an operation against scam centers in Phnom Penh, Cambodia, Tuesday, Sept. 22, 2026. (AP Photo/Heng Sinith)

Cambodia’s Scam Crackdown Has Failed to Stop the Industry’s Global Spread

 at the United Nations in Geneva to decide whether to negotiate a treaty on autonomous weapons. They must act decisively: machines of this kind have already 

 in Zaporizhzhia last July. As defense companies 

 to bolt AI onto their products, many more could die without new rules governing this technology.

, when autonomous drones killed humans — probably for the first time. A swarm of ten machines slammed into a field near Bakhmut, ending the lives of at least two combatants. Ukrainian drone maker Alexander Kokhanovskyy said the targets were identified by AI: “We just launch it and we know everything will be dead” in the selected area. This lethal targeting without human intervention announces a new regime of violence that we are utterly unprepared for: although international laws govern human decision-making, they fail to capture the new realities of digital warfare.

. Given their ability to adapt in real time, machine learning systems are responsive to countless environmental stimuli. To achieve their goals, they can identify and exploit opportunities that their designers did not foresee, exhibiting what scientists describe as “

” behavior. Put simply, their choices cannot be fully predicted or explained ahead of time as 

. Kill decisions are also deeply networked, arising from sensor and platform interaction rather than from a single, isolated judgment. In contrast, international humanitarian law assumes unity of purpose by treating each attack as flowing from a discrete decision; yet new technologies like drone swarms are decentralized by definition. Not surprisingly, the International Committee of the Red Cross has 

 these unpredictable systems to be prohibited outright.

Some argue that drone commanders are liable under existing laws for damage caused by their weapons. According to 

, operators deploying the system remain responsible for any fallout, including its precautionary assessments. But using a legal framework that relies on individual intent is difficult when human agency is severed from the drone’s decision matrix. For example, assigning blame to humans becomes highly questionable when equipment is operating behind enemy lines, where drones fly in dark mode and pilot input is completely absent: in these electronically jammed missions, it becomes impossible to verify what the drones saw or targeted — even after strikes happen — and human control is reduced to an illusion.

Even if commanders could be blamed for crimes committed with these weapons, current rules need updating for a second, more fundamental reason: humans have become so entangled with technology in waging war that they have surrendered many cognitive functions to machines. Think of the infrastructure enabling the remote war on terror, where drone pilot contributions to critical command decisions became 

 in cybernetic assemblages. This symbiosis is not slowing down: a 2019 US Army report 

 neural, ocular, and auditory enhancement as technically feasible by mid-century, further blurring the lines between operators and the platforms they deploy. In a world where cybernetic systems are generating violence exceeding human intent, the concept of a distinct “human soldier” making a deliberate choice is quickly becoming an obsolete legal fiction.

If policymakers keep clinging to this fiction, they risk ignoring a widening enforcement gap because international law barely registers the new paradigm of degraded human control. Consider the 

, when a Russian drone guided by experimental AI killed three Ukrainians. The unit, resembling a small model airplane, was programmed by human operators to destroy propane fuel tanks at a gas station. Forensic analysis showed that the drone was flying with no links to a human operator and used a commercially available Nvidia chip to steer itself. After failing to clear a building, it exploded near sheltering civilians. Who intended those deaths? The drone operators selected a target category, not a victim.

To strengthen accountability, officials meeting in Switzerland next week should start negotiating a binding treaty. The 

 on the table drafted by governmental experts offers a good start, but fails to close the enforcement gap for weapons systems operating in unobserved environments.

The report’s language covers human control before and during deployment. But once a drone is jammed, only inputs made prior to launch remain relevant for liability. This could leave operators off the hook for violations because targets are not “seen” before a drone goes dark: AI models are trained to learn what a target category looks like, but are never shown the actual asset. Such an attack could still violate humanitarian law, which 

 to minimize collateral damage. But given the absence of human input in final target selection, accountability would be elusive. Similarly, international criminal law 

 tied to a specific attack: a commander who launches a system that kills unforeseeably in an unobserved window is unlikely to be prosecuted. And worryingly, the text assigns no responsibility to the tech companies building these systems, who fiercely protect their commercial secrets from public scrutiny.

This new algorithmic reality urgently needs new rules that reach beyond individual human intent: enforceable obligations for states that field unpredictable killing machines, and liability for the companies that build them. If diplomats fail to negotiate a new treaty, ever more lives will hang by a line of code.

 of Nvidia CEO Jensen Huang in connection with extramarital affairs with staff and one involving an alleged Chinese spy will likely increase tensions between the US and China. On the US side, it will almost certainly lead legislators and administration regulators to tighten controls on Chinese access to chips and chip manufacturing equipment and might reopen the debate on a US ban of Chinese AI models. But it is not likely to scuttle the AI safety dialogue scheduled to be part of the conversation next week in advance of the meeting between Presidents Donald Trump and Xi Jinping in Shenzhen.

In strict logic, the Huang resignation should not undermine the force of the arguments in favor of a relaxed export control policy, considerations that Carl Schonander and I outlined in an 

 for Brookings over the summer. US export controls are a policy of the past. China’s authorities do not want US chips in their AI stack because it creates a vulnerability the US can exploit anytime it wants. They are encouraging local chip suppliers, principally Huawei, and understand that for a time there will be a capability gap with Nvidia chips for training AI models. But better that, they feel, than permanent dependence on the US. All export controls do is facilitate the Chinese government’s push to self-sufficiency.

But regardless of the merits of the policy debate, the Huang scandal has allowed a bipartisan policy consensus to reassert itself that the US should tighten the screws on Chinese AI companies. Former officials in the Biden administration are pushing for increased controls. Nicholas Burns, former ambassador to China, 

 on CNBC that Congress should push for a “complete ban” on the sale of advanced semiconductors into China. Chris McGuire, who was a leader in tightening export controls on China in the Biden administration, 

 strong regulations to ban “Chinese companies from renting remote access to the AI chips that they are currently prohibited from purchasing.”

 to tighten tech controls on exports to Chinese companies are 

 of the National Defense Authorization Act, including the 

, which would give Congress power over advanced chip exports, the 

, which would require businesses to track their chips to combat smuggling, and the 

Multilateral Alignment of Technology Controls on Hardware (MATCH) Act

, which ⁠would push US allies to restrict more chipmaking equipment to China.

The House of Representatives could very well accept these measures as part of agreeing to pass the NDAA. Current Minority leader Hakeem Jeffries (D-N.Y.) 

 the measures in the NDAA that restrict tech exports to China. As 

 does the current House Speaker Mike Johnson (R-La.). In addition, the Senate could take up the House-passed 

, which would give the Commerce Department further authority to close the remote server loophole that allows Chinese companies to rent restricted chips at off-shore data centers.

Even if Congress fails to act, it seems likely the administration eventually will. Reports 

 that last week Trump authorized a crackdown on companies suspected of diverting Nvidia's advanced AI chips to China. Despite Commerce Secretary Howard Lutnick’s reported reservations, this week, the Commerce Department 

 it is finalizing new rules to block access by Chinese companies to restricted chips housed in servers outside China in Malaysia, Singapore, Thailand and several other listed countries.

Huang’s downfall will also reignite the debate over 

 defending open-weight models, even when there was reportedly a push in Washington to ban them.

What do these developments mean for the AI safety dialogue between the US and China? In September, US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng agreed to set up a formal ⁠dialogue on AI, including an incident line for communications about AI safety incidents. Follow-up technical discussions are scheduled before the next meeting between Xi and Trump on Nov. 18-19 in Shenzhen.

This AI dialogue could provide a forum for the exchange of views on the risks and benefits of cutting-edge AI models, as well as a discussion of AI risks and how to mitigate them.

China seems prepared to move beyond its past focus on content regulation of AI systems to seriously address cyber risks, risks of developing biological weapons and the need to keep AI systems under full human control, so that they reliably do what people want them to do and only what people want them to do. The World Internet Conference’s Specialized Committee on AI, an international group affiliated with the Cyberspace Administration of China, the nation’s de facto AI regulator, has just released a 

 on Frontier AI Safety and Governance that uses Western-style vocabulary and threat assessments for understanding and evaluating AI risks.

The conventional wisdom is that the Huang scandal will halt any further progress on AI safety discussions between the US and China. As Bloomberg columnist Catherine Thorbecke, 

, “This scandal broke mere weeks before US-China AI safety talks were set to take place, extinguishing already faint hopes for cooperation on that front.”

But that is not how US Treasury Secretary Scott Bessent apparently sees it, and he is leading the US side of the talks. Speaking at the National Committee on US-China Relations Gala on Nov. 11, he 

 that despite the Huang scandal he still wants to use the meeting between US and China next week to advance the AI safety dialogue. He is even putting more on the table than a simple communications channel. Instead, he expects representatives from the two countries "to exchange best practices for evaluating and testing advanced models before their release, and for keeping human judgment firmly in command of this technology.”

As US policymakers ratchet up pressure on China not just on tech exports but generally, it is increasingly unlikely that the Presidential meeting next week will produce any breakthroughs on trade or a movement to make the current period of stability a permanent feature of the bilateral relationship. But these negative developments can be isolated from the technical concerns involved in AI safety. AI safety involves common threats that neither country can respond to effectively on its own. A rogue AI system or a highly capable AI model in the hands of a dangerous non-state actor pose risks to the critical infrastructure of both countries. The good news appears to be that despite an inevitable increase in tensions in connection with the Huang affair, cooperation to make progress on AI safety appears to be on track.

The author is a policy advisor at European Digital Rights (EDRi).

published its long-awaited Digital Fairness Act (DFA), accompanied by proposals

 to strengthen cross-border consumer protection enforcement and simplify certain existing consumer rules. Together, they could significantly change how the EU regulates commercial practices and, ultimately, 

how extractive business models negatively impact fundamental rights

. A gargantuan task, even if one worth tackling.

replace the Unfair Commercial Practices Directive (UCPD) with a directly applicable Regulation

. Most importantly, it introduces new prohibitions against manipulative design, addresses addictive features and tackles problems ranging from influencer marketing and gaming practices to subscriptions, personalized pricing and AI agents acting on people's behalf.

There are several things to welcome, and this Commission team deserves credit for years of serious work. Its 

 estimates that the proposed measures could prevent €13.7 billion in consumer detriment annually, including the value of time lost. This adds to the 

negative impact on fundamental rights that has been evidenced for years.

The European Digital Rights (EDRi) has been pushing for stronger protections against manipulative and exploitative commercial practices for years, including 

the need to modernize consumer law to better protect fundamental rights online

. I won't attempt a comprehensive assessment of every provision here, given that several deserve their own analysis, particularly the rules on AI agents, gaming, unfair pricing and the accompanying enforcement reforms. This piece concentrates on something running through the proposal that is difficult to ignore: how much responsibility the law continues to place on people to protect themselves against practices that companies deliberately design, test and profit from.

One of the most promising parts of the DFA is its new blacklist of ten digital deceptive design practices, the so-called “dark patterns”: fake notifications, repeated attempts to make people reverse their decisions, deliberately complicated refusal mechanisms and other familiar tricks would be prohibited outright. This is welcome, even if we've spent years debating whether existing consumer law already prohibits deceptive design. Often it does, but that hasn't stopped these practices from spreading. Forcing regulators to establish unfairness through complex legal assessments hasn't exactly delivered spectacular results for protecting people from such practices.

There is also a broader prohibition on manipulative interface design. 

Its relationship with existing consumer law and the Digital Services Act

 particularly given the overlapping rules and enforcement authorities. Still, 

: clear boundaries for practices that companies should have stopped deploying years ago. This, however, makes the Commission's hesitation elsewhere even more difficult to understand. If we can prohibit some manipulative design practices outright, why is the EU so reluctant to take a similar approach to certain (also manipulative) addictive features?

Are we really doing cookie banners again?

Take Article 17, which explicitly recognizes that features such as autoplay, uninterrupted scrolling, engagement streaks and certain AI interactions can encourage compulsive or excessive use. For consumer law, this is an important development. Businesses increasingly profit from keeping people engaged, including when no purchase takes place. The proposal even recognizes that continuing to use a service can constitute a commercial decision.

And what does the law require companies to do about addictive features? Largely, provide an off switch. We know from the documents that the Commission actually considered requiring attention-maximizing features to be disabled by default for everyone. Its impact assessment explains that this option was rejected partly because of concerns about disrupting business models and consumer experience.

We should find that reasoning rather difficult to accept. Of course people enjoy engaging services and personalized recommendations. But enjoying something tells us very little about whether we want companies continuously testing how to keep us online beyond our intentions. And the Commission's own consultation findings complicate the argument: among respondents supporting EU action on addictive design, 

58% favored allowing consumers to opt into such features

 rather than being exposed to them automatically.

There is also a too familiar political risk: we've spent years watching companies turn cookie consent from a perfectly valid legal and protective basis into an obstacle course, only for the resulting frustration to become an argument against EU privacy and data protection regulation. 

Now imagine another set of inconvenient controls, confusing settings and repeated choices: people will inevitably become frustrated, industry will complain about Brussels ruining the user experience, and 

eventually someone will suggest weakening the rules (they will call it ‘simplification,’ though)

. I'd rather we didn't repeat that exercise.

There is a legal question here, too: the text prohibits both specific manipulative interface design and unfair commercial practices in general. Could some of the addictive features covered by Article 17 already be unlawful under those provisions, even when an opt-out is available? That relationship needs clarification, because providing an off switch should never become a convenient defense for manipulation.

Apparently, being vulnerable is no longer enough

Consumer law has traditionally distinguished between the average consumer and particularly vulnerable consumers: when a commercial practice targets particularly vulnerable people, authorities can assess its unfairness from the perspective of those people, rather than the average consumer. However, 

, increasingly sophisticated commercial systems make that distinction harder to sustain.

Vulnerability in the digital realm is situational: financial pressure, exhaustion, emotional distress or loneliness may change how someone responds to commercial pressure. Companies observe people's behavior, infer their weaknesses and adapt their practices accordingly. Even someone who ordinarily makes informed decisions can still be manipulated by a system continuously optimized to exploit predictable behavioral responses.

 the limitations of traditional consumer concepts. But something interesting happened between the leaked draft and the final proposal: the former referred broadly to consumers who were particularly vulnerable because of disability, age, credulity or personal circumstances; the official proposal instead refers to a specific misfortune or circumstance serious enough to impair someone's judgment.

That considerably narrows the circumstances covered by this particular protection, potentially making it harder to assess commercial practices from the perspective of people experiencing temporary or situational vulnerabilities.

Of course, the general prohibition of unfair commercial practices still applies to everyone. But it’s difficult to understand why the Commission chose a more restrictive formulation when its own research documents how digital businesses exploit behavioral vulnerabilities. 

The EU should preserve stronger protection for people facing particular vulnerabilities, while recognizing that these can arise temporarily, affect anyone and sometimes result from the commercial practices themselves. That also means requiring companies to identify and prevent foreseeable exploitation, rather than expecting people to demonstrate that their personal circumstances were serious enough to deserve protection.

The problem becomes especially apparent with personalized practices. The proposal restricts certain harmful personalized pricing practices when traders know users are minors or particularly vulnerable. Yet companies increasingly infer and exploit vulnerabilities that may be temporary, difficult to categorize, or deliberately generated by the commercial environment itself. All of the above does not even take into account the potential for even more data collection and processing, and thus mass surveillance, of it all.

Why should consumer law require such a high threshold for vulnerability when identifying and exploiting more ordinary weaknesses has become a commercial capability (one that many companies are proud of)? Perhaps we should spend less time asking whether people are vulnerable enough to deserve special protection, and more time preventing businesses from creating or exploiting those vulnerabilities.

‘Fairness by design,’ but who's responsible for proving it?

The Commission's impact assessment discusses introducing a 

, something many civil society organizations have supported. The legislative proposal contains useful elements, including the prohibition of manipulative interface design and the existing requirement that companies act with the care and skill reasonably expected of a responsible business. But it stops short of establishing a clear horizontal duty requiring companies to assess, document and prevent foreseeable manipulation across their commercial systems.

While we welcome that the text would help courts and authorities obtain evidence about how online commercial practices operate and affect consumers, this still places the burden of proving unfair practices on regulators, civil society and affected consumers rather than companies who hold the information. Companies know which designs they tested, what they optimized for and how people responded. They know which interventions increased spending, reduced cancellations or encouraged longer engagement. Regulators, consumers and civil society organizations, hand in hand with the impacted people whose rights have been violated, generally have to reconstruct that information after the fact.

We know that the Commission examined stronger approaches to easing or reversing the burden of proof during its preparations. The European Parliament and Council, as co-negotiators, should revisit them. 

EDRi, for starters, would like to see proportionate obligations requiring traders to assess foreseeable risks before deploying potentially harmful commercial practices, alongside documentation requirements explaining their design choices. For certain practices already known to pose serious risks, the law could go further: presume them unfair unless the company can demonstrate otherwise. That doesn't mean imposing identical duties on a small online shop and a company operating sophisticated behavioral optimization systems, because the law, based on the proportionality principle, can distinguish between them.

But if companies can measure exactly how their systems influence us, asking them to demonstrate that their commercial practices are fair hardly seems unreasonable.

Children deserve protection. So does everyone else

The relationship between the DFA and the proposed EU KIDS Act (which is 

already threatening to make the internet less safe

) adds another layer of difficulty. The two proposals establish different protections depending on age, including stronger restrictions for minors on certain personalized advertising and gaming practices, alongside separate safeguards against addictive design.

There are welcome limits in the DFA on requiring additional personal data processing to identify minors. Nevertheless, the combined approach may encourage further age classification and potentially intrusive and harmful verification. And the underlying question remains: why should protection against manipulative commercial practices depend so heavily on someone’s characteristics, including age in this case?

Children may face particular risks, but harmful commercial practices don't suddenly become harmless when someone turns 18. And yes, comparisons with alcohol, tobacco or gambling sometimes appear in this debate. Those sectors show that consumer choice alone is not always an adequate response to harmful commercial practices. In addition to that, digital spaces are also places where people (including children) communicate, organize, learn and exercise fundamental rights. Importing approaches centered on restricting access would have serious consequences.

The more useful lesson is that foreseeable harms can justify regulating what companies do, instead of concentrating on who should be excluded from their services. We can draw red lines around harmful commercial practices instead. That means companies delivering a 

, something that would not only relieve people and their rights, but also create a more level playing field amongst companies. That would address the source of the problem without making intrusive age checks the condition for enjoying basic protections.

European Commission's recent preliminary findings concerning TikTok 

under the Digital Services Act make this particularly relevant. Authorities are thus already examining whether engagement-driven content recommender systems create unacceptable risks, and we believe that the DFA could complement that work through clearer rules across the digital economy, including beyond the biggest platforms. Because 

we should be addressing the commercial incentives and optimization practices 

that encourage exploitation, rather than treating each harmful feature as an isolated problem.

Let us get this straight: the European Commission recognizes addictive design, yet largely relies on opt-outs; it studies situational vulnerability, yet narrows the relevant legal language; it examines fairness by design and stronger evidentiary obligations, without clearly translating them into preventive responsibilities. What does this mean for the next steps?

Civil society groups want to be optimistic: for once, Brussels has presented legislation intended to strengthen digital rights and, given 

, we'll take that as a reason to hope for the better, even if only shortly. The European Parliament and Member States now have an opportunity to strengthen the proposal and give it actual teeth. Safer defaults, clearer prohibitions, meaningful remedies and proportionate corporate accountability are badly needed and would go a long way.

For years, people have been expected to read privacy notices, manage consent, adjust settings and defend themselves against commercial systems designed by companies with vastly greater information and resources. Companies design these systems: they test their effects, understand their commercial objectives and profit from their operation. How much longer should people have to 

 to change course on AI. Voters elected new leaders for federal and state offices that they believe will favor people over Big Tech and take a more muscular approach to regulating AI and data centers. The popular demand for intervention is now unmistakable. But what kind of intervention is necessary?

New alliances are popping up to try to answer that question. For instance, a new “

” of more than 60 organizations launched just weeks before the midterms. It joins 

 already underway to return technology toward the path of augmentation — using technology to make people more capable, rather than replacing them.

But first in social media and now in AI, those efforts have mostly been 

, fragmented, little-recognized, and under-resourced. I have been 

 for an alliance like this for some time, after seeking technology that 

Now there is new hope for realignment, spurred by widespread fear that we are at a tipping point. For those who want technology to be pro-human, it is time to treat the new problems of AI and the still unresolved problems of social media as related parts of the same challenge: building a digital ecosystem that serves people and democracy, not corporate platforms.

This is one of the great challenges of our age. Over the past decade of social media, as it became clear that the technology was 

 human thinking and collective decision-making rather than augmenting it, many have come to understand that this is not just a technology problem, but a broader 

 one. There are many technical and nontechnical levers to reverse it, but no silver bullet. I see three overarching disconnects that a coalition could help fix — and offer a few simple framing principles to address them.

Disconnect 1: AI and social media are merging

The first disconnect is that AI is not separate from social media, but is fusing with and absorbing it. The so-called “AI” and “social media” fields 

. AI already shapes what we see on social media, it is starting to take over some of social media’s role in filtering information, and it draws on social media content (which is very uneven in quality) for training. So far, the hyperscaling of AI has largely overlooked the fact that humans evolved and flourished as social creatures — 

 of social media’s failure to serve real human connection and 

. At the same time, the platforms we call “social media” are becoming increasingly less social, suggesting a need to face a new reality of "

But can humanity survive and flourish in a “post-social” world of people talking to computers, rather than one of people talking to people, augmented by computers? To channel RFK Sr., who borrowed from George Bernard Shaw, media scholars must help us see things as they are, and ask why, and media architects must dream of things that can be better and ask, why not? For AI to become 

 it must help people connect with each other. Think of this as “

” (AI and social media as one), or as “

I hope the Pro-Human AI Coalition will also address the inseparable cause of pro-social media. That would seem consistent with the 

 from Project Liberty (a founder of the coalition) and the Social Web Foundation (PL-SWF), which makes a clear case for that holistic focus, surveying key issues raised as AI and social media merge. Whatever one’s assessment of 

 is already a clear and present danger common to both AI and social media.

Disconnect 2: Tech is ignoring our social nature

The second disconnect is that neither mainstream AI nor social media are being built to support the rich social dynamics essential to humanity — the nuanced interplay between individuals and the dense network of communities and institutions that has co-evolved over millennia. Some scholars propose addressing this by combining siloed technologies that can reinforce each other (notably “

AI and the future of digital public squares

,” by Beth Goldberg and 26 coauthors; “

Improving social media for democratic discourse

,” by Fan Cheng and others). Some civil society organizations take a similarly broad view, again notably including the PL-SWF report.

Disconnect 3: Remediation requires a whole-of-society project

is that even as these ideas gain recognition, putting fixes into practice is difficult. There are an overwhelming number of them, the devil is in the many details, and broad-spectrum reform is a complex sociotechnical process gated at human speed.

 a broad, pro-human AI and social media coalition to close these disconnects: “a loosely federated, ‘big-tent’ alliance for a portfolio approach to motivate and guide the ongoing whole-of-society project needed to evolve a healthy sociotechnical ecosystem.” The newly formed Pro-Human AI Coalition promises to be the start of that — and can help guide a new awakening in US policy.

My suggested high-level architecture includes emerging support tools for social mediation processes (how others help us filter and make sense of information), individual and community control of attention, community representation, and a relentless focus on whose interests AI serves. This architecture offers a positive vision that goes beyond fixing problems. It would empower people and the collective intelligence of their many overlapping communities to select, govern, build, interconnect, and evolve the agents and services that shape how we think together, so that they serve everyone.

This Part 1 provides an overview of a few guiding principles. Part 2 expands on how they function to support pro-human objectives and then concludes with a call for a “big tent” alliance to obtain agreement on — and guide ongoing development toward — those shared principles. A 

 provides additional notes, references, and updates.

Focusing on a few big-tent framing principles

These are wicked problems with no simple solution. They are far too complex, nuanced, and context-dependent for any one person or group to fully understand. Additionally, progress has been slow because dominant platforms’ business models and centralized designs underweight the public good and user rights, while policy responses have remained limited. What can shape and hold a diverse portfolio of efforts together is a small set of shared principles that work at different levels: 

 and a simple, essential and overarching question: 

 These principles are aligned with the more detailed 

, but meant to be simpler and more universally compelling.

Debates rage over expression as a matter of “free speech rights” in isolation, often on a “for me but not for thee” basis. But the real issue is the breakdown of 

 of how social creatures think. Online media harms are commonly viewed as expression management problems, but it is managing attention where we have lost our way. Attention had been organically manageable because social influences helped guide it. Now, faced with more content creation than anyone can process, 

we need feed algorithms, or “attention agents,”

 to help us. Platforms have taken over that role, but their algorithms fail to guide attention in the nuanced way human social mediation once did.

Latest

Podcast

Recent

Our content delivered to your inbox

Thank you!

Around the globe