Anthropic's moral compass architect suggested AI overcorrection could address historical injustices

An Anthropic AI ethics researcher argued in a 2023 paper that intentional discrimination in AI models could be used to combat stigmas around race and gender topics.