Leading artificial intelligence companies made a fresh commitment to developing AI safely during a mini-summit in Seoul. Google, Meta, OpenAI, and other industry giants pledged to shut down their advanced systems if they cannot manage the most extreme risks.
Key Outcomes:
- Voluntary Safety Commitments: Major AI companies, including Amazon, Microsoft, Samsung, IBM, and others, agreed to ensure the safety of their most advanced AI models. They promised accountable governance and public transparency.
- Global Network of Safety Institutes: World leaders from 10 countries and the European Union agreed to establish a network of publicly backed safety institutes. This network will advance AI research and testing, building on institutes set up by the UK, US, Japan, and Singapore since the November AI Safety Summit at Bletchley Park.
- Universal Guardrails and Dialogue: U.N. Secretary-General Antonio Guterres emphasized the need for universal AI safety measures and continuous dialogue. He warned against a future where AI power is controlled by a few or by algorithms beyond human understanding.
Safety Frameworks: The AI companies committed to publishing safety frameworks outlining how they will assess risks. In extreme cases where risks are deemed “intolerable,” these companies agreed to implement a “kill switch” to halt the development or deployment of their AI models.
Focus on Key Issues: Aiden Gomez, CEO of Cohere, highlighted the industry’s focus on pressing concerns like misinformation, data security, bias, and maintaining human oversight. He stressed the importance of prioritizing efforts on the most likely risks.
Global Efforts: Governments worldwide are racing to regulate AI as the technology rapidly evolves. AI’s potential to transform various aspects of life—from education to privacy—has prompted urgent action. The U.N. General Assembly approved its first resolution on AI safety, and the European Union’s AI Act is set to take effect later this year. Additionally, the U.S. and China recently held high-level talks on AI.
Summit Details: The Seoul summit, co-hosted by South Korea and the UK, included a two-day meeting where industry leaders and international organizations discussed AI safety. South Korean President Yoon Suk Yeol and British Prime Minister Rishi Sunak participated in virtual sessions with other world leaders.
The summit expanded its agenda to cover innovation and inclusivity alongside safety, reflecting a balanced approach to AI’s potential risks and benefits. The outcomes of these discussions aim to contribute to a safer and more inclusive AI future.
As AI continues to advance, these commitments and global collaborations are crucial steps towards managing its risks and harnessing its benefits for humanity.
