At the Artificial Intelligence Summit held in Seoul on Tuesday 21 and Wednesday 22 May, 16 major companies and 27 countries, in addition to representatives of the European Union, signed two declarations pledging to improve the safety of artificial intelligence systems on the market and under development.
Forced to implement risk monitoring and mitigation systems by increasingly consistent legislation, the 16 major signatory companies, including Amazon, Google and OpenAI, declare that they “undertake to develop and deploy their AI models and systems responsibly”. They announce that they will “set out thresholds at which severe risks posed by an [AI] model or system would be deemed intolerable” and undertake not to develop or deploy these models and systems if the risks are indeed too high.
In addition, around 20 countries, including France, the United States, South Korea and the United Kingdom, as well as the EU, have signed a declaration to jointly develop shared risk thresholds and have committed to investing in partnerships with the private sector and the research community to decide on the thresholds to be put in place.
The issue of safety linked to the large-scale development of generative AI is becoming an increasingly important part of the international debate, particularly around election time and in view of the consequences that a lack of control could have.
The signatory countries also affirm their common intention “to guide the design, development, deployment, and use of AI in a manner which harnesses its benefits for good”.
See the statement of the 16 companies: https://aeur.eu/f/ccb
See the EU27 statement: https://aeur.eu/f/ccc (Original version in French by Isalia Stieffatre)