Most AI governance frameworks describe what should happen. Four components, three operating models, and one constitutional wall describe what holds when capability exceeds control. The dividing line: if no named human stands at the checkpoint with binding authority, the framework is Responsible AI regardless of what the policy calls it.

Open source at https://basilpuglisi.com/haia-human-artificial-intelligence-assistant/

#AIGovernance #CheckpointGovernance #OpenSource #ResponsibleAI #AIassisted

The Register: Welsh government used Copilot for review to justify closing organization. “The Welsh government used Microsoft’s Copilot to help write a review of an industry liaison body that it then scrapped, its chairman has told a Senedd committee. The government set up Industry Wales as a state-owned company to run sector forums for aerospace, automotive, and technology in 2013, adding a […]

Michigan Advance: Michigan’s use of AI to process SNAP applications draws concerns about past automation failures. “The Michigan Department of Health and Human Services has begun using artificial intelligence to help boost the number of Supplemental Nutrition Assistance Program cases it can review, a department official told members of the Senate Appropriations Subcommittee on DHHS last week.”

AFP: OpenAI shelves plans for erotic chatbot. “OpenAI has put plans for a sexually explicit chatbot on hold indefinitely, the company said Thursday, amid mounting concerns about the societal and reputational risks of releasing such a product.”

Idaho Education News: ‘The genie’s out of the bottle:’ Little signs artificial intelligence education bill. “Idaho Gov. Brad Little signed a bill instructing the Idaho Department of Education to develop a statewide framework for integrating generative artificial intelligence (AI) into classrooms.”

Gizmodo: Attorney Hit With Historic Fine for Citing AI-Generated Cases. “A court in Oregon has issued a fine of $10,000 to an attorney who submitted a legal brief with citations and quotes hallucinated by AI, according to a new report from the Oregonian. It’s the highest fine yet for citing fake cases in the state and would have been higher, but the judges offered some leniency, according to […]

The Register: Anthropic tweaks timed usage limits to discourage Claude demand during peak hours. “Anthropic on Wednesday adjusted its opaque usage limits for Claude customers by reducing the power of the services it delivers during times of peak demand, in an effort to balance demand with its capacity to deliver service.”

PsyPost: Most Americans don’t fear an AI apocalypse, according to new research. “An online study examining attitudes towards AI indicated that most people tend to hold optimistic views toward AI and tend to disagree with extreme negative attitudes. Moreover, people with higher social health, higher agreeableness, lower neuroticism and loneliness, as well as those more familiar with […]

Search Engine Roundtable: Google AI Mode Tests Changing Some Links To Overlay Cards – Reducing Clicks. “Google is testing changing out some of the links within AI Mode to show as overlay cards, instead of take you directly to the website it is mentioning. This obviously will result in fewer direct clicks to those websites.”

Ars Technica: Claude Code can now take over your computer to complete tasks. “The company has announced that Claude Code (and its more casual user-oriented Claude Cowork) can now ‘point, click, and navigate what’s on your screen’ to ‘open files, use the browser, and run dev tools automatically’ when necessary to complete tasks.”