Anthropic received a C+ in a recent safety ranking, the highest score among nine major AI firms, yet the entire industry struggles with existential risks.
TL;DR
- Anthropic scored highest with a C+ in AI safety rankings.
- No company achieved an A, indicating widespread safety issues.
- The report highlights inadequate measures against existential threats.
What happened
The Future of Life Institute released a semi-annual AI safety ranking on July 7, 2023.
Anthropic topped the list with a C+, while Meta improved to fourth and xAI fell to seventh.
Mistral, included for the first time, ranked last, citing the report's framework as unsuitable for its open model approach.
Why it matters
The low scores across the board raise concerns about the AI industry's commitment to safety amid growing capabilities.
Investors and developers may need to reconsider partnerships with companies that fail to prioritize safety and governance.
The report's findings could influence regulatory scrutiny and public perception of AI technologies.
Key facts
- Anthropic received the highest overall score of C+ in AI safety rankings.
- No company scored an A in any category, highlighting industry shortcomings.
- Mistral ranked last in the report, marking its first appearance in the rankings.
- Meta moved up two spots to fourth place, while xAI dropped three to seventh.
- The ranking evaluated companies based on six categories, including risk assessment and governance.
- Anthropic has faced criticism for its military engagements despite previously banning such uses.
Context
The AI safety landscape is increasingly scrutinized as companies develop more powerful models.
Concerns about existential risks, such as the pursuit of artificial general intelligence (AGI), are becoming more pronounced.
With the rise of military applications for AI technologies, the need for robust safety frameworks is more critical than ever.