Social media firms will face fines of up to €35 million by the end of this year if they continue to allow their AI tools to be used to create non-consensual sexual images of men, women and children online.
MEPs have voted in favour of the new rules at the European Parliament in Strasbourg, as part of the wider Digital/AI omnibus bill by a significant 423 to 57 majority, with 174 abstentions.
The new rules will now be formally approved by the European Council on 29 June.
That bill specifically prohibits social media firms from making available to customers systems which are used to create non-consensual intimate images of men, women and children.
Under the new rules, if a social media firm ignores this position after December 2026, it will face fines of up to €35m, or up to 7% of its total worldwide annual turnover for the previous financial year.
The ‘nudifier’ section of the Digital/AI omnibus bill was drawn up by a number of MEPs and officials, including Renew Group member and Independent MEP for Ireland South Michael McNamara, who is also the parliament’s co-rapporteur for its Civil Liberties, Justice and home Affairs committee.
Mr McNamara said: “This omnibus was about establishing legal certainty by extending certain timelines while preserving the AI Act’s architecture and strengthening protections where possible.
“We agreed to a limited change on machinery products, with clear safeguards, and secured an outright ban on AI nudification apps.
“They [nudifier apps] impact real people, overwhelmingly women, with the purpose of humiliating, degrading and objectifying them. I’m proud that this parliament fought for the ban, which will enter into force before the end of this year.”
Other measures contained in the Digitial/AI omnibus bill include:
– AI-generated content or images must include a water mark confirming it is not real, a plan which was initially due to be introduced earlier in the year before being included in the omnibus bill
– A total EU ban on any system a social media provider may develop which can in theory be used to create non-consensual and sexually explicit images, unless the system includes “adequate” and provable technical safeguards to prevent the creation of the images
– A clearer definition of what some social media firms describe as “safety components” which ensure AI systems are only used to assist a person or improve information access rather than other uses
– The extension to some “small mid-cap enterprises” of existing small to medium business exemptions which allow them to continue to use AI tools if they are proven to be necessary in supporting the firm’s development
The new rules will be approved by the European Council on 29 June before coming into effect on 2 December.
