From the abstract:
This paper critically analyzes the governance arrangements of OpenAI and Anthropic. These firms have an unusual built-in conflict: each raises funds from profit-seeking investors, then lets self-appointed individuals decide how much profit to sacrifice for the firm's mission-ensuring its AI benefits humanity.