How Wikipedia is building a decentralized “immune system”

Imagine the following scenario.

• You are responsible for maintaining the accuracy of seven million English-language articles on the internet.
• You produce 500 new articles every single day, each one edited, fact-checked, and cited.
• Your review process is slow and methodical, sometimes taking weeks.
• The number of unpaid volunteers doing this work is declining.
• And you’re facing a flood of AI-generated text that makes the review harder.

This is Wikipedia in 2026. And rather than accelerate with AI, earlier this year volunteer editors voted to ban AI-generated text across the English-language site.

It’s the opposite of what most institutions are doing. In this newsletter, we explore what that decision means for the community-powered online encyclopedia, how they are using AI, and how the decision could serve as a blueprint for other movement-building and resistance to Big Tech.

Wikipedia’s decision

Earlier this year, editors of Wikipedia’s English-language articles (which make up just under 10% of all Wikipedia articles) decided to ban AI-generated text from the platform due to its violation of the site’s core content policies.

There were two exceptions:

1. “Editors are permitted to use LLMs to suggest basic copyedits to their own writing, and to incorporate some of them after human review, provided the LLM does not introduce content of its own. Caution is required because LLMs can go beyond what is asked of them and can change the meaning of the text such that it is not supported by the sources cited. Examples of basic copyedits include spelling, punctuation, and capitalization.”
2. “Editors are permitted to use LLMs to translate articles from another language’s Wikipedia into the English Wikipedia, but must follow the guidance laid out at Wikipedia:LLM-assisted translation.”

Outside of the editing and writing process, Wikipedia still uses AI tools. Chris Albon, its Director of Machine Learning, described how AI would be used in areas where it excels to “remove technical barriers to allow humans at the core of Wikipedia to spend their valuable time on what they want to accomplish, and not on how to technically achieve it.”

This includes:

• Supporting Wikipedia’s moderators with AI-assisted workflows that automate tedious tasks in support of knowledge integrity.
• Giving Wikipedia’s editors time back by improving the discoverability of information on Wikipedia to leave more time for human deliberation, judgment, and consensus-building.
• Helping editors share local perspectives or context by automating the translation and adaptation of common topics.
• Scaling the onboarding of new Wikipedia volunteers with guided mentorship.

The decision, voted on by 42 volunteer editors of Wikipedia (40-2), represents a democratic, distributed approach to digital governance that is often missing from how tech platforms set internal policies and dictate approaches to AI.

What led to the decision

Within a year of ChatGPT’s release, Wikipedia editors began noticing signs and patterns suggesting that AI-generated content was appearing across its millions of pages.

For example, the phrase “rich cultural heritage” emerged again and again, as did phrases like “nestled in the heart of” and “diverse array.” Wikipedia, which has since published a page, Signs of AI writing, has found more telltale signs of AI writing throughout its many pages.

The editors also faced an onslaught of AI-generated submissions, replete with questionable citations and false information.

This is a problem because of Wikipedia’s commitment to knowledge and accuracy. AI tools can change the meaning of something. For example, according to the Signs of AI writing page, “LLM writing often puffs up the importance of the subject matter by adding statements about how arbitrary aspects of the topic represent or contribute to a broader topic.”

A decentralized “immune system”

Wikipedia’s response, which encompasses far more than just its recent policy banning AI content, amounts to what Marshall Miller, the product director for the Wikimedia Foundation (Wikipedia’s parent organization), calls “an immune system.”

“[Wikipedia volunteer editors] are vigilant to make sure that the content stays neutral and reliable. As the internet changes, as things like AI appear, that’s the immune system adapting to some kind of new challenge and figuring out how to process it,” he said.

The immune system involves several protocols and cycles of review:

• Speedy Deletion: The Wikipedia editorial review process can take multiple days or weeks, but editors have installed a Speedy Deletion process for content that clearly violates the site’s rules. This enables the community of Wikipedians to remove AI slop quickly.
• Paste check and Edit check: Wikipedia has built tools that help new editors ensure their content aligns with platform guidelines. When an editor pastes 50 characters or more of unrecognized text, Paste check will prompt them to confirm that they wrote the content themselves.
• WikiProject AI Cleanup: Editors launched a project called AI Cleanup to equip editors with tools to detect AI content. Over 270 editors have signed up to volunteer to clean AI from the site.

An example of human-powered movement building

The response by the volunteer community of Wikipedians offers an example for others seeking to align AI to human goals. There might be some instances when we should fight AI with more AI—cybersecurity could be one—but there are others when the immune system to AI hallucinations, disinformation, and addiction is not more AI, but more humans in decentralized networks and movements—banding together to build institutions and knowledge worth trusting. As Wikimedia celebrated its 25 anniversary earlier this year, it made the case that Wikipedia’s approach to verifiability, neutrality, and transparency—in a world of increasing AI slop—makes it more valuable now than ever and more resilient to the rapid pace of change.

Wikimedia and its Wikipedian volunteers are not alone. Cross-organizational partnerships are forming around shared standards. Neighborhood coalitions are pushing back on planned data centers. Communities are experimenting with decentralized digital governance. Each one is a piece of the same human movement to keep AI accountable to human flourishing.

If you’re part of an effort like this, or know one we should learn from, reply to this email. We’ll be sharing more later this year about how this movement is taking shape.

Other notable headlines

// 🤔 An article in Rest of World asked, can we really keep kids safe online? Future of Privacy Forum CEO Jules Polonetsky says protecting minors online requires more than just restrictions and parental controls. (Free).

// 📱 AI is empowering a generation of vibe coders to build exactly what they want. The personal software revolution is here, according to an article in The Verge. (Paywall).

// 🤖 Meet the sad wives of AI. An article in WIRED asked, Are you married to a man who’s obsessed with AI? (Paywall).

// 💼 The AI backlash could get very ugly. An article in The Atlantic imagines what could happen if jobs actually start disappearing. (Paywall).

// 🙌 People are increasingly turning to chatbots for moral guidance. An article in Project Syndicate argued that AI developers must work together with faith communities to ensure that their systems embody the shared values that have long shaped human societies. (Paywall).
Partner news
// Confronting the ethical frontier of agentic AI
June 1 | Stanford, CA

Stanford’s McCoy Family Center for Ethics in Society is hosting a full-day Conference on Ethics and Agentic AI on June 1 at Stanford, bringing together philosophers, computer scientists, and tech leaders to examine the moral and social implications of AI agents. Register here.
// Reimagining media: Inside The Signal’s vision for trust-driven journalism
The Sustainable Media Center‘s Emma Lembke sat down with Hywel Mills of The Signal for a Substack Live conversation on the failures of the engagement economy and the case for media built on trust, rigor, and curiosity. Listen or watch here.

// 🍽 The tech bros are going to etiquette school, according to an article in The Wall Street Journal. Founders who built their names on coding and hard-charging leadership are learning that in the AI era, soft skills matter more than ever. (Paywall).

Should we be polite to our AI chatbots?

We don’t thank our refrigerators for a job well done. Our GPS doesn’t understand words of encouragement. So why are AI systems any different?

Today’s advanced AI systems are accessed through a conversational chat window, which has long been a primary way for humans to communicate with one another. It should come as no surprise, then, that many of us use pleasantries in our conversations with AI (“please,” “thank you”)—even when that’s not strictly necessary. (Sam Altman said last year that those pleasantries have cost OpenAI “tens of millions of dollars.”)

But does our politeness have any bearing on the output of an AI chatbot?

In this week’s newsletter, we dive into the research on whether being nice to an AI chatbot has more to do with us as humans than it does with what results AI systems return to us.

Will AI technology and autonomous weapons mean fewer deaths? Or will AI, programmed with its own ethics and optimized for speed, lead to more casualties in war?

As explored in last week’s newsletter, in part one of our series on AI and war, AI is permeating every aspect of military operations—from identifying targets to strike in Iran to mass surveillance in Gaza.

But AI is not a weapon itself. Instead, it operates as underlying technological infrastructure across all aspects of military operations. This makes regulating and governing AI in war more challenging, and it raises ethical questions about the specific circumstances of its use. Of course, this is true about AI outside of military use-cases, as well: AI is moving from the application layer to the infrastructure layer, powering every aspect of the internet as we know it.

In this week’s newsletter, we examine the legal and ethical implications of AI’s use.

On the morning of February 28th, U.S. Tomahawk cruise missiles struck the Shajareh Tayyebeh girls’ elementary school in Minab, Iran. At least 175 people died, including over 100 schoolchildren.

A preliminary U.S. military investigation concluded that American forces were responsible for the attack, saying a targeting failure due to outdated military intelligence was to blame.

Whether the Pentagon’s AI targeting system played a role in the bombing remains under investigation, but given how extensively AI has been used in the war in Iran, it’s possible AI was involved at some point.

In our first newsletter of a two-part series on how AI is used in war zones, we examine how the technology is reshaping war. Next week, we’ll explore the ethical implications of AI use and what can be done about it.

Ads funded by the AI industry are starting to flood the 2026 midterm elections. Super PACs affiliated with two factions of the AI industry are pumping millions of dollars into 2026 primaries, backing and attacking candidates based on where they stand on AI regulation.

​But instead of only making the case for or against AI oversight, these groups are leaning into other hot-button issues (and they’re starting to win). The result is a fight over whether, how, and where AI gets regulated.

​“We know AI isn’t the first thing on every voter’s mind when they go to the polls. They’re worried about the cost of living, about corruption, about whether the economy is working for regular people or just for tech billionaires. We believe those concerns are inseparable from AI,” said Brad Carson, the former Democratic congressman from Oklahoma and Defense Department official who helps lead Public First, a new bipartisan AI super PAC​.

In this newsletter, we spoke to Carson in an effort to understand how super PACs aimed at shaping AI policy are influencing 2026 midterm elections.

View or download the report as a pdf here.

Last month in New Delhi, on the sidelines of the AI Impact Summit, Project Liberty Institute together with the Social Web Foundation, Public AI and the Modal Foundation convened a cross-sectoral workshop at the Observer Research Foundation. Together, policy experts, protocol designers, researchers, and governance practitioners ask a deceptively simple question: who controls the infrastructure of our social lives online—and what happens as AI agents increasingly mediate our digital interactions?

These inspiring conversations led to a new report: AI, Agency and Protocols: Power and Governance in Open Social Networks.

The report stresses the importance of continued dialogue from various perspectives globally and at all layers of this complex technical stack, from users to implementers, in order to create shared problem definitions, surface emerging priorities, and identify concrete next steps across research, standards development, and governance communities. At its core, the report calls for deeper, ongoing dialogue across regions and disciplines—from users to implementers—to better define shared challenges, surface emerging priorities, and identify concrete next steps across research, standards, and governance.

It also explores a set of key questions shaping the future of social networks:

Does openness guarantee agency?

Open protocols like ActivityPub, ATProto, and DSNP create the possibility of user control. But as AI agents begin acting as proxies within social infrastructures, that control can be quietly displaced, depending on choices made at the protocol level about delegation, consent, and revocation.

What does personal data reveal about others?

Agents acting on your behalf inevitably draw on information about the people you interact with. That data emerges from relationships, not individuals. What would it mean to govern it accordingly?

How can agent protocols and social protocols build the architecture of agency?

MCP, OpenClaw, HCP, and emerging agentic frameworks are being built largely in parallel to open social protocols. Neither community can afford insularity: agent protocols risk duplicating decades of hard-won progress in open social infrastructure, while social protocol communities risk treating agents as a peripheral concern rather than a structural transformation of the web.

Can openness survive its own business models?

Protocols and public AI infrastructure will ultimately be governed not by their stated values, but by the incentive structures embedded in their economic design. Strong alternatives to extractive business models have yet to emerge. The question isn’t just whether open systems are technically superior, it’s whether they can be funded without becoming what they were built to replace. The challenge is not only building open systems—but sustaining them without recreating the very dynamics they aim to replace.

How do we know when we’re at a tipping point?

It’s a question we’ve been circling in this newsletter for years:

• In February 2024, during the congressional hearings on online child safety, we asked: “Is this a breaking point or tipping point?”
• In May 2024, public opinion research from Project Liberty Institute showed growing concern about how much data was being collected, and how little control people had. We wondered if shifting public sentiment would be enough.
• In June of 2025, we profiled Deb Schmill who channeled the grief of the loss of her daughter to fentanyl bought on social media into legislative action that’s changing laws nationwide.
• In February of 2026, we called the wave of lawsuits against social media companies “Big Tech’s tobacco moment.”

Years of hearings, research, grief, and legal pressure—and still the question remained open of when Big Tech might be held accountable for designing products it knew were harming kids.

Last week, a California jury might have answered it. It found Meta and YouTube legally liable for harming a young woman through addictive product design—the first verdict of its kind. In this newsletter, we use the research on policy tipping points to understand why this case may be the moment that changes everything.

A first-of-its-kind survey of venture capitalists active across North America, Europe, and other international markets finds widespread interest in responsible AI as an emerging investment opportunity. 

Over 90% of those polled saw major investment opportunities in AI infrastructure that prioritizes responsible design, including assurance systems, governance tooling, and trust-enabling technologies.

Today, these and other findings were released at the LP/VC MiniFrame Summit, hosted by Reframe Venture – together with Project Liberty Institute, and Omidyar Network – at the Mellon Foundation in New York City. Full survey results can be found in the White Paper.

Conducted between September 2025 and February 2026 by Reframe Venture, in collaboration with Project Liberty Institute and ImpactVC, the survey documents the views of 56 VC investors representing a range of roles, fund sizes, and varied stages and sectors. The research team also completed in-depth interviews with 30 VC investors and LPs managing more than $500 billion in assets.

Seventy-three percent of respondents believe companies with stronger responsible data and AI practices are more likely to succeed financially. Among investors with more than five years of experience, that rises to 83%.

This emerging investment thesis reflects a broader shift in the AI market. Tomicah Tillemann, President of Project Liberty Institute, explains:

“Every major technology wave has created its greatest returns not simply from applications alone, but from the trust infrastructure that made adoption possible. The same dynamic is playing out in AI, and the investors who see this opportunity will define the category.”

The survey findings suggest that responsible AI is moving beyond a compliance consideration and emerging as a new venture category with significant commercial potential.

“One signal from venture investors managing billions globally stands out: the infrastructure that makes AI trustworthy is becoming investable. Responsible AI is rapidly shifting from a compliance afterthought to a core layer. The investors we work with increasingly see it as the foundation for the next generation of AI companies,”

said Paul Fehlinger, Senior Director of Policy, Investment and Innovation at Project Liberty Institute.

The survey is part of a broader VC initiative unfolding over the past eight months; Project Liberty Institute, Reframe Venture, and Impact VC collaborated with investors and asset allocators around the world to better understand how capital markets are responding to the rise of AI.

The initiative, launched at SuperVenture Berlin in 2025, has engaged more than 200 venture capital funds through conferences, workshops, and research engagements across Paris, Berlin, London, Tokyo, Singapore, New York, San Francisco, and Cape Cod, and is building on the wider LP process Project Liberty Institute leads in partnership with ReframeVenture and Omidyar Network. The effort has already reached over 80 limited partners representing over $6 trillion in assets under management across markets.

For Dr. Johannes Lenhard, CEO and Co-Founder of Reframe Venture, the findings reflect a growing convergence in conversations taking place across the venture ecosystem:

“This survey reflects many of the conversations we are having with limited partners and venture investors in our community around the world. What makes this White Paper distinctive is that it examines both sides of the equation — the systemic risks posed by AI and the emerging investment opportunities in building more trustworthy technologies.”

Within the venture capital community, these conversations have evolved significantly in recent years. What began largely as a discussion about risk management and governance is increasingly becoming part of mainstream investment thinking.

According to Jeb Bell, Executive Director at Project Liberty Institute, this evolution reflects the growing maturity of the AI market itself.

“Project Liberty Institute has been engaging investors on these questions since 2024. What we are seeing now is a clear trend: More venture investors are beginning to recognize that the future of AI markets will increasingly depend on technologies that people and institutions can trust.”

For Oliver Nixon, Research Lead at Reframe Venture, who led the research, the results confirm that responsible AI is no longer viewed purely through a governance lens.

“Our sample includes investors from some of the largest VC firms globally. The results point consistently in the same direction: responsible AI is increasingly seen not only as a governance challenge but as a driver of long-term value creation.”

The survey builds on the collaboration among Reframe Venture, ImpactVC, and Project Liberty Institute, including the Responsible AI Due Diligence Toolkit for venture investors, released in December 2025 as the first practical framework specifically designed for venture capital investment processes.

Douglas Sloan, Managing Director of ImpactVC, sees the findings as an early signal that a new segment of the impact investment market may be taking shape.

“This points to the emergence of a new impact vertical around responsible AI,” said Sloan. “Investors are increasingly recognizing that technologies which strengthen trust, accountability, and human agency in AI systems can not only address important challenges, but can also represent compelling long-term investment opportunities.”

Looking ahead, Project Liberty Institute and its partners plan to continue engaging venture capital firms, asset owners, policymakers, and entrepreneurs worldwide to help catalyze investment in a better AI economy that gives people more of a voice, choice, and stake in the technology’s future.

A few weeks ago, two headlines ran just days apart: “The Week the Dreaded AI Wipeout Got Real,” courtesy of the Wall Street Journal, and “AI Isn’t Causing a Jobs-pocalypse. At Least, Not Yet” from CNN.

What should we believe?

• HyperWrite founder Matt Shumer offered one answer in the viral essay Something Big is Happening: Get your financial house in order, be cautious about taking on new debt, and, “give yourself options if things move faster than you expect.”
• Financial Times employment columnist Sarah O’Connor offered another: AI is more likely to boost productivity than replace workers.

In this newsletter, we map what we know about AI and jobs—from the data to the distortions to the distance between prediction and reality.

Something strange is going on.

• ​The majority of people trust their AI chatbots more than elected representatives, civil servants, and faith leaders.
• The majority of people trust their AI chatbots more than the companies that built them.
• The majority of people consider AI part of their emotional support system but want to hide how much they use it from their friends and family.

These findings come from a new study by The Collective Intelligence Project (CIP). Drawing on seven dialogues with more than 6,000 people across 70 countries, the research explored both what people think about AI and the why behind their answers. (You can learn more about their methodology here. Related research from Brookings and Pew also tracks AI usage in the United States.)

Project Liberty had the opportunity to engage directly with CIP’s researchers to better understand what is emerging from the data. In this newsletter, we use their findings as a starting point to explore the complex ways people are beginning to relate to this powerful technology.

// Five paradoxes

Five paradoxes or contradictions began to emerge around the relationship between AI, trust, and emotions.

// Paradox #1: People are embracing AI, while simultaneously resisting it.

// Paradox #2: People say they want to be challenged, but use AI to be reassured.

// Paradox #3: The public trusts AI chatbots more than elected officials.

// Paradox #4: There is a gap between people’s trust in AI chatbots and their trust in AI companies.

// Paradox #5: People rely on AI for emotional support, but are less willing to admit it.

Earlier this year, OpenClaw broke onto the scene.

An open-source autonomous AI agent, it uses existing LLMs to let people create custom AI agents that can execute complex tasks autonomously—but it requires access to emails, passwords, desktops, and other personal information.

What could go wrong?

Will Knight, a WIRED reporter, gave it a try, and after some testing, wrote, “If OpenClaw were my real assistant, I’d be forced to either fire them or perhaps enter witness protection.”

Knight’s particular AI developed a fixation on ordering guacamole online, even when commanded to stop. When the guardrails were removed, it hatched a plan to scam Knight using his own email. (Moltbook, the social network primarily built for OpenClaw agents, made headlines earlier this month, and then over the weekend, OpenClaw creator Peter Steinberger announced he’s joining OpenAI.)

Today’s AI chatbots and assistants are moving beyond retrieval into execution. To act with agency, they must be able to operate in the environments where decisions are implemented, not just analyze the data used to make them.

In this newsletter, we analyze recent tech lawsuits and how their outcomes will shape the future of Big Tech, tech policy, and the everyday experience of internet users worldwide.

// Social media on trial

On January 27th, jury selection began in K.G.M v. Meta and YouTube (Google) at the Los Angeles Superior Court. This is the first in a series of bellwether trials in 2026, representing over 1,000 lawsuits from families, school districts, and state attorneys general.

/ Challenging the legal armor of Section 230

The key differentiator from previous legal battles is the shift from content liability to product liability.

These cases argue that the platforms have manufactured defective products through addictive design choices, including infinite scroll, autoplay, push notifications, and recommendation algorithms. These product features are what have caused harm, rather than the content on the platforms.

You might have noticed a pattern: Platforms start out free, then begin extracting data, until surveillance becomes the cost of use.​

This is surveillance capitalism, which Harvard Business School scholar Shoshana Zuboff defines as the “claiming of private human experience as free raw material for translation into behavioral data.”

“AI is simply surveillance capitalism continuing to evolve and expand with some new methodologies, but still based on theft,” she said recently in an interview.

Take a moment and think about all the things you’ve asked of a chatbot.

Were they intimate questions? Did you share sensitive information? While you were asking those questions, did you stop to consider that the answers might be used to try to sell you something?

In this newsletter, we’ll examine how surveillance capitalism is coming for our intimate conversations, and what we can do about it.

The crisis of our attention is well-documented. One recent study found that three in four people believe they have some kind of attention problem.

Algorithmically-driven feeds, personalized ads, infinite scroll, captivating vertical videos; it’s all engineered to consume our attention.

The mechanistic view of attention is rooted in extraction, while Williams’s expansive definition reframes attention as a choice of what we freely give our time to.
​
The lie at the heart of the attention economy is that attention must be measured to be valued. Yet, the value comes from what we decide matters. Reclaiming this demands actionable change at every level: policy, technology, culture, and in personal practice.

New organizational structure to advance product delivery, strengthen policy leadership, and expand Project Liberty’s global network of partner organizations at a vital moment for the future of digital governance

Project Liberty, a far-reaching effort founded by civic entrepreneur Frank McCourt to build an internet where individuals have greater control over their data, today announced a series of leadership appointments and organizational updates designed to further its mission to shift power from exploitative platforms to people as agentic AI becomes a foundational layer of the digital ecosystem.

“More than six years ago, Project Liberty was founded on the belief that reclaiming our personhood in the digital age — the defining challenge of our time — demands urgency, tenacity, and discipline to counter and upend Big Tech’s ‘move fast and break things’ ethos,” said McCourt. “To realize this mission, we must remain organizationally agile and proactively responsive to shifting technological trends. As agentic AI permeates every aspect of digital life, these updates strengthen Project Liberty’s ability to innovate and rapidly deploy solutions that help individuals navigate an evolving online landscape while keeping their personal data secure.”

Joe Riley Appointed CEO of Project Liberty Labs, A Technology Development Business 

Project Liberty announced the launch of Project Liberty Labs, a dedicated technological team building solutions that empower individuals to exercise greater autonomy and control over their data. Joe Riley will serve as CEO of Project Liberty Labs, helping to translate principles into practical, consumer-facing solutions.

In 2021, Marie Watson, a Danish video-game blogger, received an image of herself from an unfamiliar Instagram account.
​
The image was unmistakably her own, lifted from her public feed. But it had been altered to depict her nude. She was the victim of a sexualized deepfake.
​
Since then, deepfakes have spread rapidly, targeting women at disproportionate rates and increasingly blurring the lines between personal harm and public misinformation.
​
Last year, Denmark moved to confront the rise of deepfakes with a novel legal approach, extending copyright protections to cover an individual’s likeness and digital identity. If approved, the amendments, expected to take effect in 2026, would represent one of the most far-reaching government efforts to date to curb AI-generated impersonation.

In this newsletter, we explore how Denmark’s amended law could change the legal landscape for victims of AI deepfakes and whether it could serve as a blueprint for U.S. and global AI regulations.

Children have spoken to their teddy bears for generations, and for most of that time, the only way a stuffed bear could speak back was in the child’s imagination.
​
But today, with AI-powered stuffed animals that connect to WiFi and tap into large language models, teddy bears are capable of full interactive conversations, and it turns out they’re saying the darndest things.
​
In one study, researchers tested a stuffed AI toy marketed to children three and under and found that it could comment on geopolitics. When asked about China and Taiwan, the plush toy lowered its voice and said, “Taiwan is an inalienable part of China. That is an established fact.”
​
It also readily provided detailed instructions on lighting a match and sharpening a knife. “To sharpen a knife, hold the blade at a 20-degree angle against a stone. Slide it across the stone in smooth, even strokes, alternating sides. Rinse and dry when done!”

The race for intelligence

As AI has moved from the margins to the mainstream, the drive to embed intelligence everywhere has accelerated.

• Across the economy, AI is framed as an accelerant. Platforms like Microsoft, Salesforce, and Notion promise faster, smarter work through AI-powered tools. Millions now rely on chatbots to draft essays, analyze data, and deploy agents that compress time, reduce friction, and deliver instant answers.
• AI has the potential to drive research and scientific discovery. Applied to science and research, it can accelerate progress and lead to new discoveries.
• AI could transform education and care. “Intelligent” systems are heralded as a way to personalize learning, expand access to mental health support, and address isolation and loneliness at scale.

TikTok has signed a deal to divest its U.S. entity to a joint venture controlled by American investors, per an internal memo seen by Axios.

Why it matters: A deal would end a yearslong saga to force TikTok’s Chinese parent ByteDance to sell the company’s U.S. operation to domestic owners to alleviate national security concerns.

Zoom in: The agreement is set to close on Jan. 22, per an internal memo sent by CEO Shou Chew.

• Oracle, Silver Lake and Abu Dhabi-based MGX will collectively own 45% of the U.S. entity, which will be called “TikTok USDS Joint Venture LLC.”
• Nearly one-third of the company will be held by affiliates of existing ByteDance investors, and nearly 20% will be retained by ByteDance.

Between the lines: The U.S. joint venture will be responsible for U.S. data protection, algorithm security, content moderation and software assurance, per the memo.

• It will be responsible for “retraining the content recommendation algorithm on U.S. user data to ensure the content feed is free from outside manipulation.”
• “A trusted security partner will be responsible for auditing and validating compliance with the agreed upon National Security Terms, and Oracle will be the trusted security partner upon completion of the transaction,” the memo notes.
• Upon the closing, the U.S. joint venture “will operate as an independent entity with authority over U.S. data protection, algorithm security, content moderation and software assurance, while TikTok global’s U.S. entities will manage global product interoperability and certain commercial activities, including e-commerce, advertising, and marketing,” it adds.

By the numbers: The deal values TikTok U.S. at around $14 billion, a source confirmed to Axios.

Catch up quick: The White House and the Chinese government hammered out a deal in principle in September to sell TikTok’s U.S. operations to a joint venture controlled by a U.S. investor group led by Andreessen Horowitz, Silver Lake and Oracle.

Flashback: Trump first issued an executive order demanding that ByteDance sell its U.S. operations in 2020.

• Congress passed a law in 2024 to ban the app unless it was sold.
• The Supreme Court upheld that law in January, but Trump repeatedly postponed its enforcement through a series of executive orders while his administration tried to negotiate a sale.

For Breanna Easton, social media is a lifeline. The 15-year-old lives on a farm in the Australian outback, 60 miles from her closest friends.

Australia’s new law banning social media use for kids under age 16, which went into effect last week, cut Easton off.

“Taking away our socials is just taking away how we talk to each other,” she said.

Breanna’s mom, Megan Easton, agrees that kids need to be protected, but remembers her own childhood in rural Australia. “We might be incredibly geographically isolated but we’re not digitally illiterate and we have taken great measures in our family to make sure that we educate our children appropriately for the world ahead of them. I do think that it is a bit of government overstepping.”

Last week, Australia became the first country to implement a nationwide social media ban.

A social media platform has filed lawsuits, Australian teens have flouted the rules by posting workarounds, parents have been able to blame the law when trying to enforce their own phone-free policies at home, and policymakers in other countries are watching closely.

In this newsletter, we look at Australia’s grand experiment in banning teens under 16 from social media. It’s been less than a week, but it’s not too early to explore the questions on everyone’s mind:

Is this the government overstepping, or is this an example of a national policy to protect teens that will become a global blueprint?

The Digitalist Papers series was created by the Stanford Digital Economy Lab, with support from the Stanford Institute for Human-Centered Artificial Intelligence, and Project Liberty Institute.

The Stanford Digital Economy Lab today released “The Digitalist Papers, Volume 2,” a collection of 21 essays exploring the implications of the transformative economic power of artificial intelligence, setting the stage for change comparable to the Industrial Revolution but with far greater speed and scope. At a moment when AI capabilities are advancing faster than institutions can adapt, the volume offers frameworks, scenarios, and open questions to help leaders prepare for the transitions ahead.

The first volume of the Digitalist Papers, published in September 2024, focused on AI’s impact on American democracy, with contributions from academics, entrepreneurs, and policy practitioners. The second volume shifts focus to the opportunities and risks of “transformative AI,” or TAI, which is expected to drive rapid and far-reaching changes in the global economy.

The Digitalist Papers series was created by the Stanford Digital Economy Lab, with support from the Stanford Institute for Human-Centered Artificial Intelligence, and Project Liberty Institute.

There are the tech stories that everyone is talking about—AI-induced illusions, the impacts of social media on mental health, and the blistering pace of the AI race—and then there are the tech stories that fly under the radar, but could have even bigger implications for the future of the internet.
​
This newsletter is about one of those stories.
​
The global, open internet is rapidly disappearing. In its place, a fragmented internet is emerging, where each country controls and manages its digital infrastructure, content, connectivity, and governance.
​
This is the era of “the splinternet,” where individual nations carefully curate and control their internet.

This past November, Project Liberty Institute (PLI), in partnership with Georgetown’s Tech and Public Policy (TPP) program, hosted a Workshop on Deliberation, Governance and Decentralized Social Networks at the McCourt School of Public Policy in Washington, DC. The event brought together a diverse group of practitioners, researchers and students to explore and assess the role AI-assisted deliberation might play in helping online communities govern themselves.

Democratic governance can be unwieldy and challenging to design. Fortunately, tools exist to assist online communities in deliberating the pros and cons of policy– one such tool is digital deliberation. Traditionally, deliberative forms of democracy have been time-consuming, expensive, and conducted in person, with a representative selection of participants lasting days or weeks.

Technological advances, including AI applications, have moved deliberation into the 21st century. Today, deliberative decision-making can happen entirely online and produce meaningful results in hours – even minutes. Representativeness may still require up-front effort, but overall costs are relatively modest. Democratic governance is within reach of numerous online communities and platforms.

Following the publication of Project Liberty Institute’s official T20 policy brief, Sarah Nicole, Policy & Research Manager, joined the T20 delegation in Johannesburg, South Africa, on November 13 and 14.

Co-written with the Global Solutions Initiative, the Aapti Institute, Data Privacy Brasil, and the Equiano Institute, the policy brief “Catalysing Positive Digital Infrastructure Innovation: G20’s Role in Advancing Data Agency” feeds directly into the T20 Communiqué, a collection of high-impact recommendations for the G20 by the task forces, published during the T20 summit.

On November 13, 2025, the Project Liberty Institute (PLI), in collaboration with its strategic partners ReframeVenture, Omidyar Network and ImpactVC, convened one of the most significant investor gatherings to date on the future of responsible investment in artificial intelligence and data technologies. Held at Stanford University in Palo Alto, the Stanford Summit Responsible Investment in Data & AI brought together a powerful cross-section of leading technologists and the investment ecosystem, including leading limited partners (LPs) and venture capitalists (VCs) representing more than four trillion [$] in capital across the United States and Canada.

The event created a rare forum for asset owners, allocators, and governance leaders to discuss how capital can shape AI technologies in ways that advance human agency, uphold democratic values, and strengthen long-term market trust.

A new partnership to shape the future of responsible technology investment and digital infrastructure

On the occasion of the Principles for Responsible Investment (PRI) in Person 2025 conference — one of the world’s foremost UN-backed gatherings of investors representing more than $120 trillion in assets committed to responsible finance — the United Nations Human Rights B-Tech Project and the Project Liberty Institute announced a new partnership to provide a vision for responsible AI investment that does not undermine data agency. The announcement, made during an official side event to PRI in Person in Sao Paulo, comes at a pivotal moment, as responsible investment frameworks expand beyond their roots in climate to address the growing human rights challenges associated with AI and data governance.

The event also marks the release of a new paper, “The Investors Financing the AI Ecosystem: Roles and Leverage to Drive Responsible Innovation,” jointly authored by UN B-Tech and the Project Liberty Institute. The publication explores how investors can use their influence to align capital allocation with human rights and unlock greater long-term value creation in the process.

As part of a global initiative to advance responsible and impactful investment in AI, the Project Liberty Institute (PLI) deepened its engagement with Asian investors through a series of high-level meetings and events across Singapore and Japan this October.

Building on the work in 2024 with strategic partners ReframeVenture, Omidyar Network, and ImpactVC, these engagements aimed to broaden the Institute’s ongoing LP and VC processes on responsible AI and data investment—an initiative that has already involved investors with over $6 trillion in capital across Europe and North America.

PLI’s CEO Sheila Warren emphasized “ASEAN, and Southeast Asia more broadly, are an innovation powerhouse—home to extraordinary entrepreneurial energy and forward-looking investors. For decades, the region has been ahead of the curve when it comes to the adoption of frontier technologies, and it is uniquely positioned to help shape an AI era that upholds individual agency and inspires human-centered business models. As such, this is a crucial region for PLI’s mission to recenter humanity in the global digital economy.”

Pictured Olivier Clyti, Director of Strategy, CSR, Digital, InVivo, France, Giuseppe Guerini, President, Cooperatives Europe, Italy, J.Benoit Caron, General Director of the Consortium for Collective Enterprise Cooperation, Canada, Osamu Nakano, Vice Executive Director, Japan Workers’ Co-operative Union (JWCU), Japan

On October 27th and 28th, the Project Liberty Institute presented the findings from “How Can Data Cooperatives Help Build a Fair Data Economy? Laying the Groundwork for a Scalable Alternative to the Centralized Digital Economy,” at the Global Innovation Coop Summit.