AI Auditors and Independent Assessment Standards
California and numerous countries are advocating for mandatory audits by independent evaluators to enhance the safety of artificial intelligence systems and prevent potential risks.
To mitigate the risks posed by artificial intelligence technologies, California and dozens of countries worldwide have brought mandatory audits by independent AI evaluators to their agendas.
AI Audit Laws in California
Governor Gavin Newsom signed laws establishing standards for independent verification organizations and creating an auditor registry.
Signed last month, these laws rank among the significant steps taken to increase the reliability of artificial intelligence systems.
International Support and Calls
According to the Ministry for Foreign Affairs of Finland, an international joint declaration calling for independent AI assessments has gained support from nearly 30 countries.
More than one hundred researchers and experts signed a letter advocating for the establishment of standards for independent auditors.
Potential Conflicts of Interest and Risks
The prospect of independent auditors signing contracts with the technology companies whose products they audit brings the risk of conflicts of interest.
Experts warn against repeating mistakes similar to the bond rating system seen in the 2008 financial crisis.
Financing Model Debates
Instead of companies paying auditors directly, it is proposed that funds be collected in a shared pool.
This idea is also supported by AI labs such as Anthropic to preserve the impartiality of the independent audit process.