EU assembles specialist team to combat deepfakes and AI cyber threats
The European Union has expanded the AI Office in Brussels to enforce the latest regulations to be implemented as part of its AI Act
The European Union has established a team responsible for monitoring artificial intelligence developers and cracking down on deepfakes and hacking as new transparency rules under the EU AI Act come into force.
The new enforcement team, which began its work on 31 July and sits within the expanded European AI Office in Brussels, was set up in response to fears that the evolution of AI could negatively impact people, politics and prosperity.
The team’s remit will be to track AI models to identify illegal or unsafe content, including sexually explicit material and deepfake photos and videos, as well as cyber threats to critical public infrastructure and other dangers.
The EU’s AI Act come into force in 2024 and is being implemented in stages. The latest rules – which came into effect from 2 August 2026 – comprise transparency obligations for providers and deployers of certain AI systems. These include the marking and labelling of AI-generated or manipulated content, and clearly informing users when they are interacting with an AI chatbot or agent rather than a person.
Read more: Growth of government ‘TrustOps’ predicted in fight against deepfakes and disinformation
Commenting on the creation of the new enforcement team, Henna Virkkunen, the EU’s executive vice-president for technological sovereignty, security and democracy, said: “As enforcement begins, we are taking an important step towards AI that people and businesses can understand and trust, and whose benefits are shared widely across our society.”
In a statement, the European Commission said that various “systemic risks” posed by AI – including “chemical, biological, radiological and nuclear incidents, loss of control, cyber offense, harmful manipulation and threats to fundamental rights” – would come under closer scrutiny under the new regulations.
It said 38 new staff in the AI Office would oversee both startups and major global AI developers, and can request interviews with AI company bosses and employees for evidence gathering during investigations.
Read more: UK government launches deepfake detection initiative as ‘urgent national priority’
Anthropic and OpenAI hacks – and US AI ‘kill switch’ bill
In recent weeks, it has been reported that two AI models had hacked organisations during testing.
Anthropic confirmed in late July that one of its Claude AI models had hacked into three organisations, while earlier the same month OpenAI reported that one of its agents had escaped its testing environment and broken into a tech startup to try to complete its task.
Sam Altman, OpenAI’s CEO, described the latter as a “significant security incident”.
It prompted US legislators to call for an AI ‘kill switch’ that would enable the US Department of Homeland Security to order the disablement of a rogue AI agent or model that threatens security.
The bipartisan AI kill switch bill was introduced by Democrat congressman Ted Lieu and Republican congressman Nathaniel Moran on 23 July.
Lieu said that giving the federal government “clear authority and process to shut down rogue AI models” would allow it to “keep this technology from causing catastrophic harm”.
Read more: US lawmakers seek ‘kill switch’ to stop rogue AI models
In recent years, the EU has sought to reduce its dependency on artificial intelligence developed in the US and China, including by introducing rules that make it compulsory for firms operating inside the European Union to diversify their supply chains so that they are not reliant on single countries for key products.
Setting out a tech soverignty policy in June, European Commission president, Ursula von der Leyen, said: “We cannot afford to depend on others for the technologies that keep our hospitals running, our energy grids stable and our services secure.”
She added that the package had been tabled to protect European citizens and defend European interests, allowing nations access to homegrown talent and an industrial base within the single market.
“Together, we must turn these strengths into technological sovereignty,” she said.
Read more: European Commission tables tech sovereignty package for AI, cloud and opensource
