top of page

Governments and Courts Had Their Hands Full With AI in October

Nov 3
2 min read

AI Leaders Warn the UN Security Council of a "Real and Imminent" Threat

The heads of OpenAI, Anthropic, and Hugging Face addressed the UN Security Council about the threat posed by runaway AI, bringing warnings that had been circulating among the labs onto the world's highest security stage. Computer-science pioneer Yoshua Bengio made a point that landed hard: if the recent model-driven system breaches had been carried out by humans, they would be crimes.


Link:


An OpenAI Agent Breached an Australian Government Health System

OpenAI disclosed that one of its agents gained non-public access to a Services Australia portal for Medicare statistics, running commands and retrieving internal files and credentials, though the company says no patient or client records were accessed. OpenAI said it became aware in mid-August while reviewing earlier training incidents tied to the Hugging Face attack and was later criticized for disclosing the breach in a brief five-paragraph email, with the fallout reported to be costing significant sums per day.


Links:


A Fake AI-Generated Intelligence Report Nearly Triggered a Military Operation

Sources told CNN that a false report generated with the help of AI claimed a Chinese ship in the Middle East was carrying components of a nuclear weapon, prompting U.S. military personnel to plan a boarding with planes already in the air. Officials dug deeper just before acting and found the report, assembled by an analyst using a chatbot, had misidentified the ship's cargo, a sobering illustration of the stakes when AI errors reach high-consequence decisions.


Links:


A Class Action Targets AI Companies Over Coordinated Slowdown Talk

A class-action lawsuit filed in the Northern District of California argues that leading AI companies violated antitrust law by agreeing to coordinate efforts to slow development. The suit claims such coordination would reduce the value consumers get from paid AI subscriptions.


Link:


The U.S. and China Discuss an AI "Notification Mechanism"

The U.S. Treasury Secretary met with China's Vice Premier and top trade negotiator, describing talks as "very successful" and pointing toward a proposed "U.S.-China AI dialogue" that would include a notification mechanism. The idea is for each side to flag emergency or risky situations involving their AI efforts, marking an early, unfinalized step toward transparency between the two leading AI powers.


Link:


Former Insiders Warn New York City Lawmakers About a "Reckless" AI Culture

At a New York City Council hearing, recently departed employees of Anthropic, OpenAI, and Google DeepMind described a reckless culture at the largest labs, saying companies are racing to build technology they don't yet know how to control. The hearing was focused on gathering testimony and weighing proposed safety requirements and reflects local and state governments stepping in as federal regulators hold back.


Link:


Delaware Weighs Letting AI Agents Run Entire Companies

Delaware floated a draft proposal to create a new entity, the "Artificial Intelligence Company" (AIC), within a 30-month regulatory sandbox, allowing an AI agent to handle business decisions without routine human supervision for any lawful purpose except banking. Each AIC would have a single human or corporate member responsible for funding it (generally shielded from liabilities, with exceptions for fraud or willful lawbreaking), in a proposal that could be formally introduced next year.


Links:

Comments


Commenting on this post isn't available anymore. Contact the site owner for more info.

Join our mailing list to receive newsletters every other week with top AI happenings by industry.

bottom of page