Criticizing the unchecked development of the artificial intelligence sector, Anthropic co-founder Clark advocated regulating emergency shutdown methods under public policy in case systems become extremely dangerous.
Clark said many companies, including Anthropic, have different methods for shutting down their systems. He said the requirements and oversight mechanisms for these systems should be addressed as part of the broader policy debate over artificial intelligence.
A third-party option for an emergency shutdown button
Responding to the BBC’s question about whether companies should be required to have an emergency shutdown button and whether control of the mechanism should be handed to a third party, Clark said this was an issue that the public might want to understand and that could later be regulated.
Clark said specific statistics on the possibility of AI wiping out all of humanity would not be useful, emphasizing that the sector’s current direction needed to change.
“We are gambling by taking enormous risks,” Clark said.
Coxon accuses OpenAI and Anthropic
Jacob Coxon, who worked as a researcher at OpenAI and Anthropic for the past three years, argued that the two companies had not acted responsibly enough on artificial intelligence when he resigned from Anthropic.
Coxon alleged that the companies were engaged in an unchecked race to develop AI systems capable of improving themselves. He claimed these systems could soon become superhuman entities capable of hacking everything.
He also alleged that some industry employees developing AI models seriously believe these systems could kill humanity within 10 years.
