Category: Governance
AI governance frameworks, oversight models, accountability structures, and organisational control over AI use
Stay ahead of AI change. Get practical updates from Future Prep direct to your inbox.
By subscribing you agree to receive Future Prep news. Unsubscribe any time.
A frontier AI model that finds and exploits software vulnerabilities sits inside three legal regimes at once: NIS2, the AI Act and the CRA. For deployers that means three sets of obligations and three sets of audit risk. The fix is a single control map built now.
A new UC Davis measurement study shows AI chatbot data leakage is now the default behaviour of the modern chatbot stack. Names, emails and prompt content leave for third parties before a question is asked. For revenue leaders, this is a supply chain problem.
AI evaluation reports are landing in procurement inboxes from vendors, third parties and government bodies like CAISI. Each type answers a different question. This practitioner filter walks through three evidence types, four reading checks, the AI Act conformity routes and the four triggers that should refresh your evidence.
AI governance divergence stopped being a forward risk this week. The EU softened its AI Act, the United States normalised CAISI pre-release vetting and seven European tech CEOs called for further simplification. Three signals in seven days reset how compliance roadmaps must plan evidence, vendors and timelines.
The Digital Omnibus did not collapse on 28 April; it stopped. Until the next trilogue closes, the AI Act applies on its original schedule and 2 August 2026 is your planning deadline. Four scenarios, one quarterly plan, no parallel roadmaps.
The 28 April DMA review did not expand the law to cloud and AI. It narrowed enforcement onto two specific providers and one specific service category. The date worth pinning to your governance calendar is November 2026, not 28 April. Here is what changes for an EU mid-market stack.