Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

AI Act: Europe enforces the rules while the UN seeks a global standard

Europe's AI regulation is the world's first complete legal framework, built on four risk levels and nine banned practices. In New York, the talk turned to an international coalition for standards.

TechnologyAnalysisRachel NwosuPublished: 26 September 20265 min readSources 3
AI Act: Europe enforces the rules while the UN seeks a global standard

Regulation (EU) 2024/1689 is the world's first comprehensive legal framework devoted to artificial intelligence. The Commission presents it as the tool that makes AI systems trustworthy. Without rules, those systems would be hard to assess. When you cannot explain why a system reached a decision, it becomes difficult to establish whether someone was penalised unfairly, for example in a hiring process or in access to a public benefit.

Four levels, nine bans

The structure is based on risk and distinguishes four levels. At the highest level, the unacceptable one, the regulation bans nine practices. Among them are harmful AI-based manipulation and deception, harmful exploitation of vulnerabilities, social scoring and individual assessment of the risk of committing a crime. The regulation comes with a package of support measures: the AI Continent Action Plan, the AI Innovation Package and the AI Factories. It also brings voluntary tools such as the AI Pact and the AI Act Service Desk. These are meant to make the transition less abrupt for companies and developers.

The role of the AI Office

Enforcement rests with the European AI Office, set up within the Commission as the core of a single governance system at European level. The office plays a central role in implementing the regulation, with specific competence over general-purpose AI models, the ones on which most applications are built. This is where regulation meets engineering. Defining obligations for general-purpose models means dealing with documentation, evaluations and transparency requirements for systems that change quickly.

The international front

Globally, the road is still uphill. On 23 September, at the UN Headquarters in New York, the risks linked to the technology were discussed. OpenAI proposed leading a global coalition for artificial intelligence standards, and Chinese players such as DeepSeek reportedly took part in the discussion. The presence of competing parties at the same table shows where the game is shifting. It is no longer about the need for rules, which is now accepted, but about who writes them and with what capacity to verify them.

"A regulatory framework without enforcement capacity remains a statement of intent" is the risk flagged from several quarters.

Comments 0

Sources

3
  1. 01Commissione europea — AI Act, quadro regolatorio (Reg. UE 2024/1689)EN
  2. 02Commissione europea — European AI OfficeEN
  3. 03ANSA — OpenAI: «Usa guidino coalizione globale per gli standard dell'IA»IT

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.