- The European Parliament has voted to ban real-time, remote biometric surveillance, creating potential conflicts with EU’s 27 member nations.
- Additional regulations on artificial intelligence (AI) including GPT-4 have been passed, requiring risk assessments from companies like OpenAI and Google.
- The new AI Act could come into effect by 2026, influencing how companies utilize AI technologies.
In a landmark decision, the European Parliament has voted to implement a comprehensive ban on real-time, remote biometric surveillance, setting the stage for possible discord with the EU’s 27 member countries during forthcoming negotiations. The parliament has also endorsed additional regulatory measures on artificial intelligence (AI) technologies, including GPT-4, calling for risk assessments from companies such as OpenAI and Google.
Imposing Limits on AI
The previous consensus among lawmakers regarding a blanket ban was uncertain due to a collapsed political deal last week. The new mandate calls for constant and clear boundaries on AI use, prioritizing fundamental rights and democratic values alongside technological advancements, according to the Parliament President Roberta Metsola.
With 499 votes in favor, the AI Act was passed by the parliament, leading the way to ‘trilogue’ negotiations involving the parliament, EU member states, and the European Commission. These negotiations are intended to finalize a deal by the end of the year, potentially impacting how companies utilize AI by 2026. However, there may be attempts to apply some rules earlier, as suggested by Brando Benifei, a lead author of the act.
Contention over Public Face-Scanning
EU member states had previously agreed on allowing public facial scanning under specific law enforcement situations, a provision likely to be contentious in upcoming negotiations. Proposals to include exceptions for locating missing children and preventing terrorist attacks were predominantly rejected during the voting process.
Regulating AI and Its Use
The AI Act was initially proposed in 2021 to regulate the use of AI rather than the technology itself. It aimed at completely banning practices like social scoring and establishing standards for high-risk situations. EU member states have urged for the inclusion of general-purpose AI in the act, leading to added controls on “foundational models”, such as large language models like ChatGPT.
Future of Generative AI
EU’s decision on regulating generative AI could significantly impact the field, estimated to be worth over $1.3 trillion in the next decade. Any non-compliance with EU regulations could result in fines up to 6% of a company’s annual revenue.