Newsletters
European Union rules requiring companies to identify AI-generated and manipulated material took effect Sunday as part of the bloc's Artificial Intelligence Act.
Under Article 50 of the law, providers of chatbots and other interactive AI systems must tell people when they are communicating with AI, unless that fact is already obvious.
Providers of systems that produce synthetic text, images, video or audio must also embed machine-readable markers allowing the material to be detected as AI-generated or manipulated, according to the European Commission.
Companies and other professional users deploying AI systems must apply clear, perceivable labels when publishing deepfakes or AI-generated text on matters of public interest that have not undergone meaningful human review.
They must also notify people exposed to AI-based emotion recognition or biometric categorization systems.
Deepfakes are defined under the law as AI-generated or manipulated images, audio or video that resemble existing people, objects, places or events and falsely appear authentic.
AI-generated text on politics, public health, security, the environment, consumer safety, science and culture must be labeled when published without human editorial oversight.
Substantive human review removes the labeling requirement, though the commission noted that checks limited to spelling, grammar or formatting do not qualify as editorial control.
Artistic, creative, satirical and fictional works containing deepfakes are subject to special provisions: disclosure is still required but can be presented in a way that does not interfere with the work's display or enjoyment.
Individuals using AI for personal, non-professional purposes fall outside the law's scope; those using it regularly for economic gain or professional activity may be classified as deployers and covered by the requirements.
The rules also apply to companies based outside the EU when their AI output is used within the bloc.
Generative AI systems already on the market before Aug. 2 have until Dec. 2 to meet the machine-readable marking requirement, though other transparency obligations are not subject to this grace period. Content produced before Aug. 2 does not require retroactive labeling.
National market surveillance authorities will lead enforcement, alongside the EU's AI Office in areas under its supervision.
Violations can bring fines of up to €15 million ($17.3 million) or 3% of a company's worldwide annual turnover.
The commission has also approved a voluntary code of practice giving companies a standard route to demonstrate compliance; about 190 companies and organizations had signed on by the end of July.
Several major technology platforms introduced AI labels before the requirements took effect. TikTok says more than 3 billion videos have been tagged through creator disclosures, detection systems and technologies including Content Credentials and invisible watermarking.
Meta applies "AI Info" labels to content on Facebook and Instagram when it detects industry-standard signals or receives user disclosures.
Google has signed the EU's transparency code and is developing marking and detection systems using SynthID and Content Credentials, though the company has said overlapping labels and legal notices could create confusion rather than clarity for users.
Set as preferred source
Set as preferred source
1 min read
1 min read
