The European Fee would require tech platforms like TikTok, X and Fb to detect synthetic intelligence (AI)-generated content material to safeguard upcoming European elections from misinformation.
The fee has initiated a public session on proposed election safety pointers for very massive on-line platforms (VLOPs) and really massive on-line engines like google (VLOSEs). The suggestions intention to scale back democratic threats posed by generative AI and deepfakes.
The draft pointers supply examples of potential measures to deal with election-related dangers, together with particular actions associated to generative AI content material, planning for threat mitigation earlier than or after an electoral occasion, and offering clear steerage for European Parliament elections.
Generative AI can mislead voters or manipulate electoral processes by creating and disseminating inauthentic, deceptive artificial content material concerning political actors, false depiction of occasions, election polls, contexts or narratives.
Guaranteeing election integrity means defending public discourse – a lot of which happens on-line.
We’re in search of views on our draft Digital Companies Act pointers to deal with the rising on-line dangers to electoral processes whereas safeguarding the constructing blocks of our democracies ↓
— European Fee (@EU_Commission) February 8, 2024
Presently open for public session within the European Union till March 7, the draft election safety pointers advocate alerting customers on related platforms to potential inaccuracies in content material produced by generative AI.
Based on the draft, the rules additionally recommend guiding customers to authoritative info sources and state that tech giants ought to implement safeguards to stop the technology of deceptive content material with important potential to influence person conduct.
On AI-generated textual content, the present advice is for VLOPs/VLOSEs to “point out, the place potential, within the outputs generated the concrete sources of the knowledge used as enter knowledge to allow customers to confirm the reliability and additional contextualize the knowledge.”
The draft steerage’s proposed “greatest practices” for threat mitigation draw inspiration from the EU’s just lately approved legislative proposal, the AI Act, and its non-binding counterpart, the AI Pact.
Associated: The EU’s Artificial Intelligence Act, explained
Worries about superior AI methods, akin to massive language fashions, have emerged because the viral surge of generative AI in 2023, making instruments like OpenAI’s ChatGPT broadly identified.
The fee didn’t say when precisely corporations shall be compelled to label manipulated content material beneath the EU’s content material moderation law, the Digital Companies Act.
Nonetheless, in an organization weblog put up, Meta introduced plans to introduce fresh guidelines concerning AI-generated content on Fb, Instagram and Threads within the coming months. Any content material acknowledged as AI-generated, whether or not by way of metadata or intentional watermarking, will obtain a visual label.
Journal: AI has killed the industry’: EasyTranslate boss on adapting to change
