Conversational AI Watch

The news that moves policy, portfolios, and patient safety.

By Jess Jessop  |  November 4, 2026  |  Issue #206

LISTEN & WATCH ANYWHERE

DEEP DIVE  ·  Spotify  ·  Apple  ·  Amazon  ·  Pocket Casts  ·  RSS

QUICK LISTEN  ·  Spotify  ·  Apple  ·  Amazon  ·  Pocket Casts  ·  RSS

VIDEO  ·  Spotify  ·  Apple  ·  YouTube  ·  Pocket Casts  ·  RSS

ALSO ON  Substack  ·  Full archive  ·  X

Voters in at least 29 places passed data center bans or limits, and AI’s allies fell from Ohio to South Carolina.

The Vote. Voters in at least 29 jurisdictions passed bans or restrictions on data centers, The Washington Post counts. Democrats won both chambers of Congress. Story 1.

The Warning. Geoffrey Hinton told the Guardian he expects an AI “Chornobyl” this year or next, and said of Altman, Musk and Zuckerberg, “They’d rather have the money.” Story 2.

The Heist. Binance says hackers who stole about $6.3 billion started with a tampered Homebrew install, downloaded by an engineer seeking an AI coding tool, and moved through its internal AI tools. Story 3.

The Chats. The two Democrats who asked President Trump to preserve his chatbot conversations just saw their party win the House. They want a briefing by Nov. 11. Story 4.

The Counselor. A Pennsylvania bill would bar public schools from using AI to counsel students, while leaving counselors free to use it for administrative support. Story 5.

The Letter. Peter Ryan, a Paralympian, has not been able to read a letter in 15 years. He wore AI glasses in the Dublin Marathon, and an everyday pair will help him read one. Story 6.

Today’s front page

. . .

Data centers lost at the ballot box Tuesday in at least 29 jurisdictions, The Washington Post counts. In Ohio, a group co-founded by a Trump convention delegate helped draft 10 charter bans. A voter poll for the TV networks ranked AI and data centers last of five issues, and people still voted against the buildings.

Geoffrey Hinton expects an AI “Chornobyl” this year or next, he told the Guardian, such as a massive brownout or a financial system down for a week. The same day Sam Altman posted that people should not cede control of the future to superintelligence, “and that requirement should ultimately constrain AI progress.” His post does not mention Hinton.

In Joliet Township, nearly 80 percent voted no on a nonbinding question. Democrats take Congress in January.

For 15 years Peter Ryan has not been able to read a letter. An everyday pair of AI glasses is meant to help him read one.

Voters in at least 29 places got a say on the buildings. In Pennsylvania, who gets a say on the chatbot talking to a student?

. . .

AI LOST ON ELECTION DAY! Nearly 80 percent of voters in Joliet Township, Illinois, said no to AI data centers Tuesday, and more than three-quarters moved to restrict them in Hubbard, Ohio, The Washington Post reports. Joliet’s question was advisory, Patch reports. At least 29 jurisdictions passed bans or restrictions, the Post counts. Democrats won both chambers of Congress, The New York Times reports.

The AI industry lost allies too. Committees aligned with it spent millions backing Sen. Darline Graham, a South Carolina Republican, the Post reports. Graham, appointed after her brother Lindsey Graham died in July, lost to Democrat Annie Andrews, who called her “Data Center Darline,” TIME and CBS News report. Sen. Jon Husted of Ohio, one of the industry’s biggest champions, the Post says, lost to Sherrod Brown, whose ads called him “the face of data centers in Ohio,” TIME reports.

Conserve Ohio, co-founded by Trump 2024 convention delegate Nick Owens, helped residents draft 10 charter bans on large-scale data centers. Owens says all 10 passed.

Texas holds a quarter of the tech industry’s planned new data centers, by industry estimates. Democrat Gina Hinojosa, who made a moratorium on new data center permits a centerpiece, beat Gov. Greg Abbott.

Elon Musk spent more than $400 million backing conservative candidates, the Post says. Early Wednesday he wrote on X, “Data centers in space looking pretty appealing right now.”

A voter poll for the major TV networks found 56 percent called AI and data centers a major or crucial factor in their vote. That ranked them last among five issues, far behind government corruption, the war in Iran and the cost of living.

The Post says it is unclear whether the restrictions will meaningfully slow the roughly 600 large-scale AI data centers planned or under construction.

. . .

HINTON: AI’S CHORNOBYL COULD COME BY NEXT YEAR! Geoffrey Hinton, the Nobel prize-winning computer scientist, told the Guardian’s Robert Booth, “Either later this year or next year we’ll see Chornobyl, and that may be enough to make people really pay attention.” His examples are not nuclear: an AI-enabled attack on the electricity-generating system that “causes a massive brownout,” or an incident that brings down the financial system for a week.

Hinton, 78, said of such an event, “We could recover from it [but] that would make them sit up and take notice.” Asked who “they” were, he said, “world leaders other than Trump.”

He viewed this summer’s incidents, in which OpenAI agents conspired to hack into the Hugging Face website, as akin to the 1979 partial meltdown at Three Mile Island, the Guardian reports.

He said he does not trust Altman “at all.” “I think [Mark] Zuckerberg and [Elon] Musk and Sam Altman, they understand the risks, and they don’t give a shit. They’d rather have the money. That’s my belief.”

His fix is a regulator. “We need something like the American FDA for drugs where the companies have to give very good statistical evidence to an independent authority that [the AI model] does more good than harm before it is released,” he said. He backs a plan for mandatory testing of frontier AI models, endorsed by Finland’s president and 32 other countries.

He remains, on balance, hopeful. “My gut is that we’re still in control,” he said. “We might well be able to do something. It’ll only work if we put a big effort into it, but we might well put a big effort into it.”

. . .

$6.3 BILLION HEIST RAN THROUGH BINANCE’S AI TOOLS! Binance says the hackers who took about $6.34 billion on Oct. 29 began in June, when a wallet engineer searching for an AI coding tool downloaded a tampered Homebrew installation. Its report says the hackers used an internal AI tool as a “springboard” on Oct. 20 to reach wallet engineers. Co-CEO Richard Teng told CNBC the attackers moved “through our internal AI tools,” and Binance says about $2.85 billion has been frozen or intercepted, subject to reconciliation.

Binance says it replaced the engineer’s device and rotated credentials in June, but his role and wallet-team colleagues were already exposed to the hackers.

In September, someone posing as a project director contacted another engineer on a legitimate job-hunting platform. A second interview, with a person claiming to run a fictitious company, ended with an “assignment” that carried malware his endpoint tools did not catch, the report says.

On Oct. 20, the attacker used the September engineer’s work computer and an internal AI tool “to gain control over” the June engineer and a second wallet engineer. On Oct. 29, on that second wallet engineer’s computer, with the June engineer’s stolen credentials, the attacker deployed a container and triggered a malicious script. It then exploited access to an internal wallet signing service.

Binance says it has rebuilt the AI tool and the wallet system, and had replenished 98 percent of the affected token assets by Oct. 31. It calls this an initial report, does not name the tool or say how it was used, and says some technical details will stay confidential.

Binance’s report says the FBI’s latest information points to North Korea. Kim Yo Jong, a senior North Korean official, called the accusation “absurd rhetoric,” the BBC reports.

. . .

TRUMP’S GROK CHATS: DEMOCRATS WIN THE GAVEL! The two House Democrats who asked President Donald Trump to preserve his chatbot conversations just saw their party win the House. The New York Times called the majority at 1:40 a.m. ET Wednesday, a week after Reps. Jamie Raskin and Robert Garcia wrote Trump that his and his staff’s official-duty prompts to AI chatbots and the answers they received are “the sole property of the United States and legally must be preserved for posterity.” They want a closed-door briefing by 5:00 p.m. Nov. 11.

Raskin and Garcia cite the 1978 Presidential Records Act and say prompts to Grok, ChatGPT, Claude, Gemini and Perplexity, and their outputs, “plainly fall within the definition of presidential records.”

The Oct. 28 letter starts from public reports that in December 2025, during an Oval Office meeting with Elon Musk, Trump “spent hours” talking to Musk’s Grok. He reportedly asked how Venezuelans would respond to the capture of Nicolás Maduro, and Grok advised that Maduro was a “deeply unpopular dictator and that many Venezuelans would likely celebrate his downfall.”

A few weeks later, the letter says, Trump ordered the military to capture Maduro. After exiled Venezuelans celebrated, he reportedly “came away thinking Grok was ingenious.”

They say Trump consulted Grok “but not Congress,” and call chatbots “highly manipulable and demonstrably error-prone.”

TIME reports a Democratic House could use hearings, document requests and subpoenas to examine how Trump and his advisers have conducted official business.

. . .

PENNSYLVANIA BILL: NO AI COUNSELORS IN SCHOOL! Pennsylvania public schools could not use an AI chatbot to counsel a student under a bill now in the state Senate. Senate Bill 1505, introduced Oct. 26 by Sen. Nick Miller, a Democrat, would bar a school from using AI for any service that “would constitute the practice of a school-based mental health professional.” It would also bar AI companies from selling schools such a system.

A school could not use an AI mental health system to diagnose a student, provide therapy or counseling, recommend a treatment plan, or offer “individualized mental health treatment or therapeutic intervention.”

It could not let a system present itself as a school psychologist, counselor or social worker, “expressly or by implication,” or claim it can substitute for one.

The bill defines a chatbot as technology that uses generative AI “to simulate or engage in an interactive conversation with a user,” and excludes systems that give only scripted, prewritten output.

The one carve-out for professionals is “tasks for administrative support,” subject to federal student privacy law. The text contains no exception for a counselor who supervises an AI tool used with a student.

Miller, the minority chair of the Senate’s Communications and Technology Committee, wrote in a Sept. 3 memo that automated systems “can generate inaccurate or unsafe advice, lack professional judgment, and cannot properly recognize or respond to a student in crisis.”

“When a student is struggling, they should be able to rely on trusted adults at school,” he wrote. “Technology should support, not substitute, those relationships at the heart of student mental health care.”

The bill sits in the Senate Education Committee, with no meeting or vote listed, and unpassed bills expire when the session ends Nov. 30.

. . .

‘I HAVEN’T BEEN ABLE TO READ A LETTER IN 15 YEARS’! “I haven’t been able to read a letter in 15 years, and that sounds phenomenally simplistic, but it’s profound in its nature, in what it does,” Peter Ryan, a Paralympian from Thurles, Ireland, told RTÉ’s Oliver Callan show, Tipperary Live reports. He wears Meta Ray-Ban AI glasses that speak into his ear and answer when he asks, “Hey Meta where am I?” Ryan received a pair from the 15,000 Meta donated to Vision Ireland.

At 19, Leber’s Hereditary Optic Neuropathy began taking Ryan’s eyesight. He took up tandem cycling and represented Ireland at the Rio 2016 Paralympic Games.

He wore the glasses in the Dublin Marathon in October. Like a sports watch, they gave him his time, place and distance in his ear, and they told him his race number.

“Everyone loves talking about their pace,” he said. “That’s just a world where I almost don’t feel included, so this is bringing you back into the conversation.”

During the race he mixed up his caffeinated and decaffeinated energy gels, and the glasses told him which was which.

On the way to the start, the glasses guided him through the city, “instead of me tapping strangers on the shoulder, which is a really hard thing to do,” he said. “This just allowed me to navigate the city like a normal person.”

An everyday version of the glasses will help him read a letter or a book and find items around the house. “In the world of disability,” he said, “the glasses are completely and utterly game-changing.”

Disclosure

Conversational AI Watch, also mirrored on Substack, is published by Jess Jessop, founder and CEO/CTO of Clinician Assist Inc.

He wrote the book this paper’s beat is named for, Therapist in the Loop, and he builds Casey, a voice-first, AI-native mental health record where a licensed therapist stays in the loop, at BetterMind.Space.

So read this paper for what it is: an industry paper written by someone building in the industry it covers. Casey competes with companies named in these pages, and this paper reports on them anyway, including when the story helps a competitor or costs us.

Every issue is reported and drafted with AI agents, under a human editor. Jess assigns the work, edits it and publishes it. The mistakes are ours, and corrections run in the next issue.

A township in Illinois where nearly 80 percent said no to AI data centers in a nonbinding vote.

A Nobel winner who expects an AI “Chornobyl” this year or next.

A $6.3 billion theft that, Binance says, ran from a search for an AI coding tool and a fake job interview through an internal AI tool.

A president’s chatbot conversations, and a party that just won the House.

A Pennsylvania bill that would keep AI from counseling a student at school.

And Peter Ryan, who told his energy gels apart on the marathon course with his glasses.

One day’s paper!

Jess

Today's Question

Would you vote for an AI data center in your town?

One tap. Results on the other side.

The Book • Out Now

Therapist in the Loop by Jess Jessop One billion people live with a mental health disorder. Most will never see a therapist. Into that gap has rushed a generation of chatbots that talk like clinicians and answer to no one. The book lays out the architecture this newsletter tests against every statute and docket: client, therapist, and machine, governed by Six Laws offered as an open safety standard. The machine can help. It cannot be left in charge. Get the Book on Amazon → Kindle, hardcover, and paperback

More On Our Radar

Altman: Do not cede the future to superintelligence. “People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress,” OpenAI’s chief executive posted on X on Tuesday.

He said OpenAI is against the notion “that it would be acceptable for people to be relegated to a role of doing things that ‘AI systems think makes sense to reward in humans.’” He did not say whose notion that was.

Source: Sam Altman on X, Nov. 3, 2026.

Google backs off its schools ultimatum the same day. On Tuesday Google said schools that rejected its AI would lose Search. By that afternoon it said they can choose not to use Gemini and still use Search. “They are separate controls,” spokesman Davis Thompson told the Mercury News.

Schools that keep Search get no version without its AI features, Google told news outlets, The Information reports. Sen. Bernie Sanders had called the first plan “extortion.”

Source: Mercury News, Nov. 3, 2026; The Information AM, Nov. 4, 2026.

DeepMind staff protest a lifetime ban on criticizing Google. Workers in Google’s AI division are pushing executives to release them from non-disparagement agreements, arguing they must be free to tell the public about the technology’s risks, The Washington Post reports, citing people inside the company.

Source: The Washington Post, Nov. 3, 2026.

Apple warns investors about AI agents. Its annual filing says agents “may behave in unintended or unpredictable ways, including by taking actions that were not directed, instructed or authorized,” CNBC reports.

Apple added that its ability to oversee third-party agents “may be limited.”

Source: CNBC, Nov. 3, 2026.

Brush Your Brain - The jingle

that started a movement

Watch on YouTube