AI Insights
Regulation, governance, strategy and tools, written for the people accountable for AI programs. Every article lists the primary sources it relies on.
Guides: AI regulation for managers · AI governance for managers · The AI Management Office · AI regulatory calendar · How AIMP compares with AIGP, AAIA and CPMAI
In March a Rome court accepted OpenAI’s challenge to a 15 million euro privacy fine. Read both halves of the story: regulators examine how models are trained, and their findings can be contested.
4 min read
Stanford’s annual report runs to more than 400 pages. Six figures from it belong in every AI strategy discussion this year.
4 min read
The framework is voluntary and free, and it has become the reference point for regulators. The useful version of it fits on two pages per system.
5 min read
The FTC’s 2023 case against a pharmacy chain is still the clearest statement of what a regulator expects before a biometric AI system touches customers.
4 min read
On March 20 the administration sent Congress four pages of recommendations, including targeted preemption of state AI laws. They are recommendations, so today nothing changes.
4 min read
Gartner expects organizations to abandon most AI projects that lack AI-ready data. “Ready” has a practical definition, and five controls that make it true.
4 min read
From 60,000 users in 2024 to about 250,000 in 2025: three design choices made the growth possible, and none of them was the choice of model.
4 min read
A year ago governments moved within days against a hosted AI app. The lesson for buyers is that “which model” is really three questions, and only one of them is about capability.
4 min read
The AI Impact Summit closed with a declaration endorsed by 88 countries and organizations. It sits beside binding laws, voluntary frameworks and targeted rules that a global company now has to read together.
4 min read
A short all-hands email set off a public backlash and two rounds of clarification. The strategy barely changed; the communication did.
4 min read
Since January 22, 2026 Korea has a comprehensive AI law with duties for “high-impact” systems and a reach beyond its borders. Fines wait for a year; the obligations do not.
4 min read
A chatbot that invented a refund policy, a legal brief with six cases that did not exist, and a hiring model that learned the past. Each failure had a known fix.
4 min read
Launched at Davos on January 22 and described as a global first, Singapore’s framework for AI agents is short, practical and usable far outside Singapore.
4 min read
Newsletter
A monthly summary of what changed in AI regulation, governance and practice, with the sources for every item. Free. Unsubscribe in one click. See a sample issue.
Next step
Enroll in the AIMP certification: $695 to enroll, 100% online, at your own pace, with a verifiable certificate when you pass the final exam.
Questions? support@iaim.ai, 9:00am to 5:00pm Pacific, Monday to Friday.
