The European Union has added 38 staff to its AI Office in Brussels as the European Commission prepares to exercise new enforcement powers over advanced artificial intelligence models from Sunday, 2 August. The expansion, reported by the Associated Press on Friday, gives the bloc more capacity to monitor global developers and investigate risks ranging from cyberattacks and harmful manipulation to threats against fundamental rights.
The change marks a major implementation stage for the EU AI Act, the risk-based law adopted in 2024. Commission guidance says its enforcement powers for obligations covering providers of the most advanced general-purpose AI models begin on 2 August 2026, after a one-year compliance period. The office already employs more than 125 lawyers, engineers, economists, policy specialists and administrators.
According to AP, the additional team will monitor companies ranging from start-ups to large American and Chinese developers. The Commission can require model documentation, request information and remedial measures, evaluate models, interview company personnel during investigations and impose sanctions. EU institutions and national market-surveillance authorities share enforcement responsibilities, with national regulators supervising many AI systems used within member states.
Transparency rules also start applying on 2 August. Commission guidance says providers must tell people when they are interacting directly with AI and must make generated or manipulated output machine-readable and detectable. Deployers must disclose deepfakes, certain AI-produced public-interest material without human editorial control, and the use of emotion-recognition or biometric-categorisation systems. A limited grace period until December applies to the marking duty for generative systems already on the market before the deadline.
The Commission identifies systemic risks from powerful models as including cyber offence, chemical or biological hazards, loss of control, harmful manipulation and damage to fundamental rights. The AI Office has launched channels through which technology workers and users can confidentially report suspected misconduct or compliance concerns, while an independent scientific panel and an EU board representing member states support consistent oversight.
The enforcement push will test whether Europe can translate its landmark legislation into practical supervision without suppressing innovation. Companies serving EU users now face closer scrutiny of technical records, risk controls and synthetic-content disclosures. Regulators will next have to coordinate investigations across 27 countries and show that requirements can be applied consistently to developers based both inside and outside the bloc.
Comments