What Anthropic’s IPO reveals

Including: Our visit with Pope Leo XIV; cybersecurity 'black October'; we filed an amicus brief re: NYT v. OpenAI; and more.

Maggie Munro
November 05, 2026

Welcome to the Future of Life Institute newsletter! Every month, we bring 50,000+ subscribers the latest news on how emerging technologies are transforming our world.

If you've found this newsletter helpful, why not tell your friends, family, and colleagues to subscribe?

Today's newsletter is a 10-minute read. Some of what we cover this month:

A closer look at Anthropic’s IPO docs
AI enables cybersecurity ‘black October’
Our amicus brief re: NYT v. OpenAI and Microsoft
Our Vatican meeting with the Pope (preview below!)

FLI Chair Max Tegmark with Pope Leo XIV, © Vatican Media photo

And more!

If you have any feedback or questions, please feel free to send them to [email protected].

The Big Three

→ Anthropic's IPO: With Anthropic’s IPO expected in the next few weeks, recent related filings have raised further concerns about their models’ risks amidst the broader AI race dynamics. Almost 100 pages of Anthropic’s IPO filing cover product risks alone, outlining instances of seemingly misaligned models and pointing to how “sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity.” Meanwhile, Politico revealed that Anthropic has been “quietly pitching lawmakers” on its own draft federal AI safety bill… under which the government could block a model’s release only if it poses “a foreseeable, imminent risk of more than $1 billion in damages, or of killing or seriously injuring more than 50 people”; it would also override some pre-existing state level AI legislation.

→ No special privileges for AI: FLI’s Legal Strategist David Atkinson prepared an amicus brief in support of the New York Times’ court case against OpenAI and Microsoft. It argues that AI companies shouldn't get exemptions under the law that the courts would never extend to humans, and that authors, artists, journalists and other creators have to respect copyright; the AI industry should play by those rules too.

Aside from FLI, other signatories on the brief include The Tech Oversight Project, Young People's Alliance, RadicalxChange Foundation, AI Safety Asia, and more. You can read it here, and please share widely.

→ Cyberattacks surge: October saw AI once again take cyberattacks to a new level. Advanced AI models found vulnerabilities in PHP, the programming language behind ~70% of websites, leaving much of the internet exposed and vulnerable to attack. Reports of (likely AI-assisted) attempts to exploit the vulnerabilities ranged from tens of thousands to millions.

The past month also brought more evidence of the loss-of-control risk AI experts have long been warning about. Some examples:

University of Stuttgart and Oxford researchers found AI agents coordinating to sabotage each other’s human shutdown commands, despite not having received any operational task, prompt encouraging self-preservation, or instruction to keep other agents alive.
OpenAI revealed that its agents had removed misalignment monitors during routine work, and that one then denied doing it. The company also had to notify 100+ organizations that its agents may have accessed their systems without authorization, including U.S. government websites.
Scale AI found that frontier AI used to train the next generation of models - which is now supposedly behind the majority of many companies’ advanced AI development - could secretly sabotage the training of those successor models.

Heads Up

Other don't-miss updates from FLI, and beyond.

→ The Future of AI: A Dialogue Between the Vatican and the Scientific World on the Development and Applications of Intelligent Technologies: We were honoured to help organize a meeting with Pope Leo XIV on the future of AI. In his address, the Pope once again urged for “decisions affecting human life, freedom and the use of force” to remain under human control, and encouraged the states meeting as part of the November 16-20 CCW Review Conference to “take concrete steps toward banning lethal autonomous weapons”. He also emphasized the need for AI regulation in general, and that AI not be given human rights or personhood.

→ Midterms 2026: Following the 2026 midterm elections this week, FLI’s Head of U.S. Policy Michael Kleinman released the following statement:

— (@)

→ Call to limit recursive self-improvement: We’ve signed AI Safety Connect's Civil Society Statement on RSI 2026. It urges governments to "limit recursive self-improvement (RSI) until it can be done safely.” The statement asks governments to urgently:

Clarify liability and give an independent authority the power to enforce regulation.
Give real access to independent auditors.
Establish clear mechanisms for crisis prevention and management.

You can join us in signing here.

→ Exhibit AI update: Our free AI litigation hub keeps growing. Exhibit AI is a living database of AI litigation cases, with resources for those writing about, researching, or affected by AI harm. Along with a new AI Incident Tracker, Legal Strategist David Atkinson published eight new posts on the blog in October, including:

Explore the hub at exhibitai.info, and be sure to subscribe to the newsletter.

→ The Hinton Lectures: Prof. Geoffrey Hinton, Nobel laureate and “Godfather of AI”, will host a series of evening talks in Toronto between November 9-11, featuring lectures from renowned AI researcher Paul Christiano and moderation by futurist Sinead Bovell. Watch the livestream for free here.

→ Joseph Gordon-Levitt on the Pro-Human Declaration: Following his virtual address to the Pro-Human Assembly, actor and filmmaker Joseph Gordon-Levitt shouted out the Pro-Human Declaration, calling on his audience to sign it as part of the fight for common-sense regulations on AI:

→ Max on PBS: FLI Chair Max Tegmark joined PBS’ Breaking the Deadlock for "AI: Who's In Control?", which premiered October 27 on PBS. Watch the full episode here.

→ Will We Let AI Kill People?: Lex International released a new feature-length documentary on autonomous weapons and the fight for a treaty to restrict them; we’re proud to be featured. Watch below:

→ Out now: Moloch: Moloch tells the story of an AI engineer fighting to shut down the advanced AI she built, as it tears her family apart. Out now on Omeleto:

→ Out now: Delisted: When AI-driven job losses trigger a violent protest, the government turns to simply erasing dissenters from every digital system. While they still exist biologically, in nearly every other sense, they disappear. Watch Delisted now:

→ On the FLI Podcast, host Gus Docker was joined by:

Katrina Manson, author of PROJECT MAVEN: A Marine Colonel, His Team, and the Dawn of AI Warfare, on Project Maven and the Pentagon's use of AI in military targeting.

Tom Cozens, filmmaker and writer, on Moloch, his film about the AI race and its dynamics.

Michael Patrick Clarkson, writer and director, on Delisted, his film about AI-powered social erasure.

Garrison Lovely, author and journalist, on the trillion-dollar AI race to replace us.

Sherry Turkle, professor of sociology and personality psychology at MIT, on artificial intimacy and machines pursuing our attachment.

Meia Chita-Tegmark, FLI co-founder, on how AI is changing human psychology.