With increasing artificial intelligence (AI) safety concerns, more than 120 leaders of global AI institutions, evaluation experts, technology organizations, and leading thinkers signed an open letter supporting independent AI evaluation that is rigorous and reliable.
The letter calls for developing global standards and shared principles to measure AI and to test industry claims of safety, accuracy, and fairness. It notes that governments and other leaders should create the conditions for credible evaluation, including making AI systems transparent and traceable, with reportable failures. Independent evaluation can help pave the way for organizations to deploy AI systems safely, with confidence and credibility.
