“I now join the ranks of former colleagues - both at OpenAI and other leading players in the industry - who have concluded that the current direction is unacceptable,” Robinson writes.
Development at OpenAI is progressing too quickly, he continues.
“As the company rushes from one launch to the next, it fails to achieve the level of caution that I believe is necessary.”
According to Robinson, AI companies must start drawing on security expertise that exists in other areas and wait for new research that ensures new models make “safe decisions even when we are not monitoring them.”
AI developers should, Robinson writes, take inspiration from the safety and risk management at a nuclear power plant or a large airport.
