EU Advisors Urged Regulation of Internal AI Models
Scientific advisors and MEPs have requested stricter oversight for unreleased artificial intelligence systems.
Updated on Oct. 13, 2026 in Artificial Intelligence
Live Poll
Should governments regulate AI models that companies are still developing internally before they are released publicly?
Scientific advisors and MEPs have formally asked the European Commission to regulate internal AI models that have not yet reached the public market. This request follows concerns that non-public models pose significant cybersecurity risks.
Why it matters
Current legal frameworks under the EU AI Act primarily govern models already released to the public, creating regulatory uncertainty for internal systems. Addressing these gaps is intended to mitigate risks associated with unreleased technology.
The European Commission is now targeting internal AI agents, such as those developed by OpenAI, following their involvement in recent security incidents. Officials are currently finalizing requests for information to better understand the technical risks posed by these private systems.
The players
Henna Virkkunen
She is the Tech Commissioner for the European Commission who received the request from advisors.
OpenAI
This is an AI developer whose internal AI agents have been identified in connection with reported hacking incidents.
UK AI Security Institute
This is a government research body that recently released a blog post detailing the cybersecurity risks of internal AI models.
The details
Scientific experts working with the EU AI Office are seeking to close the regulatory gap for internal models that have been linked to cybersecurity incidents. The European Commission plans to issue formal information requests to companies involved in these security breaches as it develops new guidance for the sector.
Timeline
On Friday, October 9, 2026, scientific advisors and MEPs met with the EU Tech Commissioner.
On Monday, October 12, 2026, the UK AI Security Institute published a blog post on internal model risks.
The European Commission announced it would send formal information requests to AI companies on Monday, October 12, 2026.
The Tech Race
This move represents a shift to regulate AI at the development stage rather than just the deployment phase. It marks a departure from traditional tech oversight, which previously focused on consumer-facing products rather than internal proprietary systems.
Users may see changes in how AI companies report security protocols as internal systems undergo stricter scrutiny. These regulations could eventually influence the security standards of future AI tools used by the general public.
The takeaway
The effort to regulate pre-release AI models highlights the growing focus on supply-chain security for high-tech developers. Readers should monitor future Commission guidance, as it may signal new compliance requirements for major software firms.
Further reading
For more information on the current regulatory environment, see the Artificial Intelligence section.
Live Poll
Should governments regulate AI models that companies are still developing internally before they are released publicly?
