-
Classify Against the 19 May Draft Guidelines Now — Before They Are Final and Before You Are Rushed
The Commission’s draft high-risk guidelines are the clearest answer yet to “is my system high-risk?” The consultation closes 23 June. Here is how to run the test this week.
-
The Explanation Gap Between GDPR Article 22 and AI Act Article 86
GDPR Articles 13, 14, 15, and 22, together with AI Act Article 86, give affected people different routes to understand and contest automated or AI-assisted decisions — but they cover different decisions, and the gap between them is where a deployer’s real obligation lives. No consequential automated decision in Europe answers to a single…
-
The Clause That Almost Vanished
The public record does not prove who wrote the Digital Omnibus. It shows something more useful: who had access, whose asks aligned with the draft, and which public-visibility safeguard nearly disappeared before the compromise text restored it in narrower form. Read the coverage of the Digital Omnibus and you come away with one fact:…
-
The Regime That Got Nothing: GPAI Enforcement Still Arrives on 2 August
The 7 May Omnibus deal — still a provisional agreement, pending formal adoption — would delay the high-risk rules. It leaves the general-purpose AI obligations exactly where they are. For model builders, the enforcement clock does not move at all.
-
Twelve Days Apart: Brussels Moved the Deadline, Then Told You How to Meet It
The EU delayed its high-risk AI obligations on 7 May and published 148 pages of classification guidance on 19 May. Read together, the two acts say the same thing — the date moved, the expectation did not.
-
The U.S. National AI Safety Board: Regulation Enters the Real World
The United States has formally established a National AI Safety Board (NAISB), an independent body modeled on the National Transportation Safety Board. Announced in early October 2025, the NAISB will investigate significant AI failures—ranging from algorithmic discrimination to catastrophic automation incidents—and publish public findings (White House, 2025). The move signals…
-
Algorithmic Power and Public Oversight: The Next Phase of AI Accountability
In October 2025, the United States announced the formation of a National AI Safety Board—a permanent oversight body modeled on the National Transportation Safety Board. Days later, the European Commission inaugurated its AI Office, and UNESCO expanded its Ethics of AI Observatory. Within weeks, three continents converged on one insight…
-
AI Security by Design: Integrating Governance into the Development Lifecycle
Integrating Governance into the Development Lifecycle Artificial-intelligence security is entering a phase where good intentions are no longer sufficient. 2025’s high-profile AI breaches—from model-prompt leaks to manipulated training datasets—exposed that most organizations still treat governance as a post-deployment activity. The new “secure-by-design” guidance from the UK’s National Cyber Security Centre…
-
Market Disclosure and the Corporate Governance Shift
A quiet revolution is taking place in corporate reporting. In their 2025 third-quarter filings, companies including Microsoft, SAP, and UBS began referencing AI risk governance alongside traditional cybersecurity and ESG disclosures (Bloomberg, 2025). These mentions are brief but significant.
-
Generative AI in Media: Deepfakes, Elections, and Authenticity Liability
As election seasons unfold across multiple continents, lawmakers and media organizations are racing to counter an explosion of AI-generated misinformation. In September 2025, the European Parliament advanced a bill requiring labeling of synthetic political content, while the U.S. Congress is considering a similar “AI Transparency in Communications Act”.