EU publishes code on GenAI transparency
The European Commission has published a final code of practice setting out how AI providers and deployers should mark and label content generated by the technology.
The commission says that the voluntary code is aimed at providing practical steps to help providers and deployers of generative AI (GenAI) meet the transparency obligations that will apply under the AI Act from 2 August.
From that date, deepfakes and AI-generated or AI-manipulated text published on matters of public interest must be clearly labelled.
Users must also be informed when they are interacting with an interactive AI system, such as a chatbot.
The code was drafted by six independent experts, with input from over 180 interested parties.
Machine-readable
It consists of two sections – one for AI providers and one for those deploying AI tools.
The code for providers sets out how to ensure that AI-generated or AI-manipulated audio, images, video, or text are marked in a machine-readable way and can be detected as artificially generated or manipulated.
For deployers, the code explains how they must clearly label deepfakes and AI-generated or AI-manipulated text published to inform the public on matters of public interest when there has been no human review or editorial control.
The commission is now inviting all providers and deployers to sign up, adding that those who do will be able to demonstrate compliance with the relevant AI Act obligations that start to apply on 2 August.
‘Show leadership’ call
The code will be complemented by guidelines that will clarify the scope of the legal obligations and address aspects not covered by the code.
Commissioner with responsibility for technology sovereignty Henna Virkkunen said that Europeans had a right to know whether what they saw, heard, or read had been made or altered by AI.
She invited AI providers and deployers to adhere to the code and “show leadership in developing and deploying AI that is responsible and trustworthy”.