Content created by an AI like ChatGPT should be © flagged Nikos Pekiaridis/IMAGO
What to do about fake news generated by artificial intelligence? The EU is now making a push.
This article is IPPEN. MEDIA in the course of a cooperation with the Europe.Table Professional Briefing – it was first published by Europe.Table on 6 June 2023.
BRUSSELS – EU Vice-President Věra Jourová sees defending freedom of expression as the EU's task. Editor's note), I don't see the right to freedom of expression for the machines," Jourová said on Monday (5 June) at the presentation of a strengthened EU code of conduct against disinformation. The signatories "should introduce technologies to recognize AI content and clearly label it for users," Jourová said.
The EU had presented the Code of Practice against Disinformation 2022. At that time, 44 companies signed the paper. These include Google, Facebook, YouTube and TikTok. In the meantime, twelve more signatories have joined. However, Twitter had left the agreement last month.
Newsletter from Table.Media
Get 30 days of free access to further exclusive information from the Table.Media Professional Briefings – the decisive factor for the decision-makers in business, science, politics, administration and NGOs.
While the code is voluntary, it is intended to be implemented under the DSA (Digital Services Act; Editor's note). The DSA requires very large online platforms such as Twitter to assess and mitigate systemic risks, including disinformation. Very large platforms that repeatedly violate the DSA risk fines of up to six percent of their global turnover.
Goal: Curb disinformation ahead of the 2024 European elections
The Commission is now tightening the code and expects companies to make greater efforts – also in view of the upcoming European elections in 2024:
- Fight against Russian disinformation, including in connection with the war of aggression against Ukraine
- more fact-checking and better funding for independent teams of fact-checkers (especially in small and medium-sized EU Member States)
- better access to background data for researchers
- Safeguards against malicious use of AI to spread disinformation
- clear labeling of AI-generated content
The Commission calls on online platforms to start labelling without delay. The signatories are expected to set out measures to label AI as early as July in the next background reports on the implementation of their commitments. Vis