Microsoft, Google, OpenAI, Mistral AI + 12 others signing AI safety standards at Seoul Summit

Sixteen leading technology companies have agreed to a set of voluntary AI safety standards at the ongoing summit in Seoul. 

These standards, established through collaboration between governments and industry leaders, outline best practices for responsible AI development. They emphasize information sharing among companies, investment in cybersecurity for AI systems, and prioritizing research into the potential societal risks associated with advanced AI.

The voluntary nature of these commitments has drawn mixed reactions. While some see them as a significant initial step towards safer AI development, others express concern about the lack of enforced regulations. Experts believe mandatory rules might be necessary to ensure companies prioritize safety measures.

The participation of companies from China (Zhipu.ai) and the United Arab Emirates (Technology Innovation Institute) is viewed as a positive development by some, suggesting a growing international consensus on the importance of AI safety.

“The Seoul summit builds upon the progress made at the Bletchley Park summit last year,” said Michelle Donovan, the UK’s Technology Secretary. “We’ve seen an increase in international cooperation, with several countries establishing their own AI safety institutes.”

Despite the potential benefits, concerns remain. 

The summit also included a closed virtual meeting between world leaders and business executives, reportedly focused on fostering further collaboration in AI safety research.

The full list of the companies to have signed up to the safety standards:

• Amazon
• Anthropic
• Cohere
• Google / Google DeepMind
• G42
• IBM
• Inflection AI
• Meta
• Microsoft
• Mistral AI
• Naver
• Open AI
• Samsung Electronics
• Technology Innovation Institute
• xAI
• Zhipu.ai

Here is OpenAI’s statement.