How can you tell if a company manages its responsible with its AI? We analyzed public data from 250+ companies and will present our conclusions.

➤ Our study

EthicsGrade collected public data about the AI ethics activities of 250+ companies in 2022🌟

We analyzed this data (using the NIST AI RMF)🌟

We especially looked for correlations between signals of good AI governance, such as having AI ethics principles, and implementation activties🌟

➤Key findings

No evidence that AI ethics principles and voluntary commitments are correlated with actually mitigating risks💻

No evidence that any of the other governance signals we studied is correlated with implementation, including thought leadership and employing specialized personnel/teams💻

Overall, more companies seem to have deteriorated than improved the volume of their AI ethics activities during 2022. Most companies stayed stable, however.💻

➤ Key insights

Many push for the adoption of AI ethics principles. For example, the US and Canada came out with initiatives to solicit such commitments. The UK thinks that voluntary guidelines can sometime replace regulation.💡

Consumers, investors, and other actors need to rely on external signals. Such actors often rely on the existence of AI ethics principles and personnel.💡

Our findings raise a red flag. Such signals may be misleading.💡

➤ The team includes , , , , and Gil RosenthalTess BuckleyDr Joshua ScarpinoLuke PattersonThorin Bristow

Hope to see you there!