Global AI Safety Under Scrutiny as Experts Call for Stronger Action
NEW YORK — A new safety ranking of the world’s leading smart technology developers has found that while some companies are making progress, the industry as a whole remains ill-prepared for the most serious risks posed by advanced systems.
The semiannual report, released Tuesday by the Future of Life Institute, a US-based think tank, evaluated nine major developers across six categories: risk assessment, current harms, safety frameworks, existential safety, governance and accountability, and information sharing.
Anthropic, a San Francisco-based lab, received the highest overall score of “C+” but no company earned an “A” in any single category. Meta moved up two spots to fourth place, while xAI dropped three spots to seventh, just ahead of China’s DeepSeek and France’s Mistral, which placed last.
Seven researchers and governance experts determined the rankings based on public data and information provided by the companies. Three Chinese developers DeepSeek, Alibaba Cloud, and Z.ai — all produced open models and landed in the bottom half of the ranking.
Open Models and Data Sharing
Mistral, included for the first time, disputed the report’s framework, saying it is not suited for its approach to developing open models which allow users to download and modify them. MIT professor Max Tegmark, president of the Future of Life Institute, expressed disappointment at Mistral’s last-place showing, noting that Europe has been a leader in safety standards.
The organization said Mistral did not respond to its survey, nor did Alibaba, xAI, or DeepSeek.
Concerns Over Military Use
The report also noted that several companies that previously banned their technology from military uses have “gradually reversed course,” including Anthropic, which the report criticized for “questionable military engagements.” According to media reports, the US government used Anthropic’s technology in military operations in Venezuela and Iran over the past year — though the company was recently banned by the Pentagon over disagreements on safety standards.
All nine companies are failing when it comes to combating “existential” threats, such as pursuing systems that reach human-level intelligence, the report said. Although “constructive attempts exist,” efforts across the board are “entirely inadequate.”
Anthropic was thrust into the spotlight earlier this year after it released its most powerful model yet, called Mythos. In early April, the company released Mythos only to a handful of trusted organizations due to its abilities to expose cyber vulnerabilities to bad actors. However, on June 12, the US government blocked Anthropic from releasing Mythos to foreigners on national security grounds. The Trump administration eventually lifted the ban on June 30.
Also Read:
GCC and Italy Sign Strategic MoU in Rome to Boost Bilateral Cooperation
How G42’s Inception42 and Microsoft Are Transforming Government Operations with AI Assistants
