World News for Students of English

14-12-2023 07:00

Europe has achieved a preliminary agreement on groundbreaking European Union regulations for artificial intelligence (AI), covering government use of AI in biometric surveillance and regulating AI systems like ChatGPT.

This political breakthrough positions the EU as a leading global standard-setter in AI legislation, marking a historic day according to European Commissioner Thierry Breton. The agreement mandates transparency obligations for foundation models, such as ChatGPT, and general-purpose AI systems (GPAI) before entering the market. These include technical documentation, compliance with EU copyright law, and detailed summaries about training content. High-impact foundation models face additional requirements, including model evaluations, risk assessments, adversarial testing, reporting incidents to the European Commission, cybersecurity assurance, and energy efficiency reports. Governments are restricted in using real-time biometric surveillance, permitted only in specific crime cases or imminent threats, such as terrorist attacks.

Difficult words: standard-setter (an organization or group that establishes and maintains standards for a particular field or industry), mandate (to require something to be done), adversarial testing (examining systems by attempting to exploit weaknesses).

You can watch the video news lower on this page.