In the News: Europe makes a Deal

The European AI "Deal"

Source:  Europe reaches a deal on the world’s first comprehensive AI rules By Kelvin Chan (Associated Press) December 9, 2023 

Summary

The European Union has achieved a landmark by securing a deal on the world's first comprehensive artificial intelligence (AI) rules. Negotiators from the European Parliament and member countries have overcome significant differences, addressing contentious issues such as generative AI and the use of face recognition surveillance by police. The resulting tentative political agreement, known as the Artificial Intelligence Act, sets the stage for legal oversight of AI technology, aiming to balance its transformative potential with concerns about human safety and privacy.

Details

• Negotiations and Agreement: The EU negotiators reached the agreement after intensive closed-door talks, spanning 22 hours in the initial session. The deal covers various aspects of AI use, including generative AI and face recognition surveillance.
• Generative AI Concerns: The recent surge in generative AI technologies, exemplified by systems like OpenAI's ChatGPT, prompted a reevaluation of the initial 2021 draft of the AI rulebook. Generative AI, while displaying remarkable capabilities in creating human-like text, images, and music, has raised fears regarding job displacement, privacy issues, copyright protection, and broader societal impacts.
• Global Regulatory Landscape: The EU has taken an early lead globally in establishing regulatory frameworks for AI. The U.S., U.K., China, and other major democracies are also developing their proposals for AI regulation.
• Foundation Models: The scope of the AI Act was expanded to include foundation models, which are critical to general-purpose AI services. These models, such as those underlying ChatGPT and Google's Bard chatbot, raised concerns due to their potential risks, including online disinformation, cyberattacks, and bioweapon creation.
• Face Recognition Surveillance: Negotiations around AI-powered face recognition surveillance systems were particularly challenging. European lawmakers sought a full ban on public use, citing privacy concerns. However, exemptions were negotiated to allow law enforcement use for serious crimes.
• Loopholes and Concerns: Civil society groups expressed reservations about the AI Act, emphasizing that the deal did not go far enough to protect individuals from potential harm caused by AI systems. Concerns include loopholes, lack of protection for AI systems in migration and border control, and the option for developers to opt-out of high-risk classification.

Reaction and Commentary

The political deal on the AI Act has been met with a mixture of approval and skepticism. While European Commissioner Thierry Breton celebrates the EU as the first continent to establish clear rules for AI use, civil society groups and industry representatives emphasize the need for detailed technical work to address the nuances of the legislation. The EU's comprehensive rules are seen as potentially influential globally, setting an example for other governments considering AI regulation.

Speculation for Future Outcomes

• Global Impact: The EU's AI rules are anticipated to have a global impact, with other countries likely to emulate aspects of the legislation in their regulatory frameworks.
• Extended Obligations: Companies subject to the EU's rules are expected to extend some obligations outside the continent, promoting a unified approach to AI development.
• Challenges and Refinements: The coming months will witness the EU Parliament's vote on the AI Act, with attention turning to addressing concerns and refining technical details. The law's full implementation is not expected until 2025 at the earliest, indicating a period of ongoing adjustments and improvements.
• Balancing Innovation and Safety: The negotiation process reflects the delicate balance between fostering AI innovation and safeguarding against potential risks. The ongoing evolution of AI technologies will necessitate continuous updates and adaptations to regulatory frameworks to keep pace with advancements and emerging challenges.
  

Immediate Impact

The immediate impact of the new European deal on comprehensive AI rules is still uncertain, as the law has not yet been finalized and implemented. However, some potential impacts include:

• Increased scrutiny of AI systems: The law will require developers of AI systems to assess and mitigate the risks of their systems, and to report serious incidents to the European Commission. This could lead to a more cautious approach to the development and deployment of AI.
• Restrictions on the use of AI: The law bans the use of AI for certain purposes, such as social scoring and real-time biometric surveillance. This could limit the potential applications of AI in some areas.
• Increased transparency: The law will require developers of AI systems to disclose information about how their systems work. This could lead to greater public understanding of AI and its potential risks.
• Global leadership: The EU is the first major jurisdiction to introduce comprehensive AI rules. This could give the EU a competitive advantage in the global AI market.

It is important to note that these are just some of the potential impacts of the new European deal on comprehensive AI rules. The actual impact of the law will depend on a number of factors, including how it is implemented and enforced.

Here are some specific examples of how the law could impact different types of AI systems:

• Facial recognition: The law bans the use of real-time facial recognition in public spaces, with some exceptions. This could make it more difficult for law enforcement agencies to use facial recognition technology.
• Chatbots: The law requires developers of chatbots to disclose information about how their systems work and to assess the risks of their systems. This could lead to more transparent and trustworthy chatbots.
• Algorithmic decision-making: The law prohibits the use of algorithms that make decisions about individuals without human oversight. This could limit the use of algorithms in areas such as hiring and loan approvals.

The new European deal on comprehensive AI rules is a significant development in the regulation of AI. It is likely to have a major impact on the development and deployment of AI in Europe and around the world.