Seven people who have worked with OpenAI said that during the OpenAI period, Dario Amodei's belief in proper AI development methods was further solidified. Daniela Amodei joined OpenAI in 2018 and later became Vice President of Security and Policy. It wasn't long before employees questioned the Amodei brothers and sisters continued to expand their influence. Three people familiar with the matter revealed that in December of that year, researcher Joshua Achiam proposed to Sam Altman that Amodei should not be appointed as Chief Ethics Officer. Achiam believes that the Amodei brothers and sisters are becoming more and more powerful, and that AI supervision work should include people with different positions. In 2018, the Amodei brothers and sisters took over the GPT project, which is a generative pre-trained converter. The six participants in the project say they are exploring a set of alignment techniques to train AI to perform tasks according to human instructions. This set of idea—proactively pushing AI to dangerous boundaries and then testing using safe methods—has continued throughout Amodei's career. OpenAI President Greg Brockman tried to participate in the GPT project, but was resisted by the Amodei brothers and sisters. Two people familiar with the matter said that Amodei has repeatedly stated to Altman that he will leave his job if Brockman is not transferred from the project. A person familiar with Altman's ideas said that Altman thought Amodei's ultimatum was too extreme and rejected his request. In December 2020, Amodei, Daniela, and their colleagues left OpenAI. Four weeks later, Anthropic was founded.

Seven people who have worked with OpenAI said that during the OpenAI period, Dario Amodei's belief in proper AI development methods was further solidified. Daniela Amodei joined OpenAI in 2018 and later became Vice President of Security and Policy. It wasn't long before employees questioned the Amodei brothers and sisters continued to expand their influence. Three people familiar with the matter revealed that in December of that year, researcher Joshua Achiam proposed to Sam Altman that Amodei should not be appointed as Chief Ethics Officer. Achiam believes that the Amodei brothers and sisters are becoming more and more powerful, and that AI supervision work should include people with different positions. In 2018, the Amodei brothers and sisters took over the GPT project, which is a generative pre-trained converter. The six participants in the project say they are exploring a set of alignment techniques to train AI to perform tasks according to human instructions. This set of idea—proactively pushing AI to dangerous boundaries and then testing using safe methods—has continued throughout Amodei's career. OpenAI President Greg Brockman tried to participate in the GPT project, but was resisted by the Amodei brothers and sisters. Two people familiar with the matter said that Amodei has repeatedly stated to Altman that he will leave his job if Brockman is not transferred from the project. A person familiar with Altman's ideas said that Altman thought Amodei's ultimatum was too extreme and rejected his request. In December 2020, Amodei, Daniela, and their colleagues left OpenAI. Four weeks later, Anthropic was founded.