What it is
The Scientific Panel of Independent Experts is an advisory body established under Article 68 of the EU AI Act, with its working details set out in a 2025 Commission implementing regulation. It is one of three advisory structures created alongside the Act’s enforcement machinery — the others being the AI Board (made up of member-state representatives) and the Advisory Forum (representing industry, civil society, and academia). Where those two bodies represent interests, the Scientific Panel is designed to represent expertise: it is staffed entirely by independent technical specialists who are not supposed to have financial or institutional ties to AI providers.
The Commission appointed up to 60 members to the panel on 1 June 2026, following an open call for applications that ran from mid-June to mid-September 2025. Members serve renewable two-year terms and are drawn from fields including frontier model research, technical auditing, systems engineering, and societal-impact analysis. The appointment process required geographic spread (at least one expert per EU/EFTA/EEA state, capped at three per state, with at least 80% of members from those regions) and attention to gender balance.
The panel’s formal powers are narrow but real: it can issue “qualified alerts” to the Commission’s AI Office when it suspects a general-purpose AI model poses a systemic risk at EU level, it can request information from model providers (this requires authorization from at least one-third of the panel’s members), and it advises the AI Office on classification criteria, evaluation methodologies, and market-surveillance priorities for general-purpose AI models.
Why it matters for AI governance and narratives
The panel is a test of a specific governance bet: that a standing body of credentialed, independent technical experts — rather than government regulators alone — can keep pace with frontier AI capabilities well enough to flag emerging systemic risks before they cause harm. The EU has described the AI Act as “the first law in the world that addresses systemic risk from AI,” and the panel is the mechanism meant to give that enforcement regime scientific legitimacy rather than purely bureaucratic authority.
For the observatory’s purposes, the panel is also a useful marker of how the “regulator” ecosystem’s framing of AI risk is evolving. Its activation on 9 October 2026 — a special meeting convened specifically to examine recent loss-of-control incidents and to prepare questions for the companies whose models were involved — signals that the Commission is willing to use this expert channel as a forward operating post for scrutinizing frontier labs directly, not just for drafting technical standards in the abstract. How labs respond to information requests coming through this channel (rather than through national regulators or informal diplomacy) is itself a data point on the balance of power between European regulators and US frontier developers.
Key facts and dates
- Legal basis: Article 68 of the EU AI Act (Regulation (EU) 2024/1689), with implementing detail in a 2025 Commission regulation on the panel’s selection and operating procedures.
- Appointment: Call for applications ran 16 June – 14 September 2025; the Commission appointed the panel’s members on 1 June 2026.
- Size and composition: Up to 60 independent experts, two-year renewable terms, with mandated geographic and gender-balance criteria.
- Special meeting, 9 October 2026: The Commission convened the panel specifically to review recent loss-of-control incidents involving frontier AI models. Per the Commission’s own announcement, the panel and the AI Office had jointly developed a set of questions to put to the companies whose models were implicated, and the panel presented safety and security recommendations to the Commission during the meeting. The Commission did not name the companies or incidents in its public announcement.
- Statement: Henna Virkkunen, the Commission’s Executive Vice-President for Tech Sovereignty, Security and Democracy, said at the meeting: “The EU has the first law in the world that addresses systemic risk from AI, and we need state-of-the-art scientific input… Europe can lead in ensuring safe and secure AI.”
- Broader enforcement context: The AI Act’s General-Purpose AI Code of Practice treats loss of control as a systemic risk category that signatories have been expected to address since August 2025, with the AI Office beginning enforcement in August 2026.
One caveat worth flagging: public reporting on the 9 October meeting does not specify which companies or which incidents triggered it. The editorial’s reference to the Commission “demanding information from OpenAI and Anthropic after cyberattacks” appears to draw on separate reporting not confirmed in the Commission’s own announcement of the scientific panel meeting; readers should treat the panel’s specific involvement in that particular episode as unconfirmed by primary sources at the time of writing.
Where to learn more
- Commission holds special meeting of Scientific Panel on frontier AI safety and risks — the Commission’s own announcement of the 9 October 2026 meeting.
- The Scientific Panel of Independent Experts: What Is It And How Does It Work? — detailed explainer on the panel’s legal basis, selection process, and powers, from the widely-cited AI Act tracker site.
- Article 68: Scientific panel of independent experts — Cambridge Commentary on EU General-Purpose AI Law — academic legal commentary on the statutory text establishing the panel.
- European Commission appoints expert bodies to support AI Act enforcement — third-party summary of the panel’s and Advisory Forum’s appointment.