- Current and former employees at AI say they are being silenced over safety issues
- One signee of an open letter claims artificial general intelligence poses great risks to humanity
- Open letter signees want a right to warn the public about the dangers of AI
An artificial intelligence (AI) expert has claimed that there is a 70% chance that AI could end humanity, and that companies developing AI are not adequately mitigating against the risks.
Stark warning
In an interview with the New York Times, Daniel Kokotajlo, a former governance researcher at Open AI, made this stark warning, adding that the risks of artificial general intelligence (AGI), which is in development, are being ignored. Instead, developers are focused on the possibilities of it, he said.
He predicted that AGI would be achieved by 2027, but that there was a 70% chance it could be catastrophic for humanity.
Kokotajlo is also one of the signees of an open letter, which says they believe in the potential for AI “to deliver unprecedented benefits to humanity.”
But the signees, all current or former Open AI and Google DeepMind employees – some of whom remained anonymous – warn that these technologies also pose serious risks. “These risks range from the further entrenchment of existing inequalities, to manipulation and misinformation, to the loss of control of autonomous AI systems, potentially resulting in human extinction.
Mitigating risks
While they note that AI companies, AI experts and governments have acknowledged the risks. They add that while they are “hopeful” the risks can be mitigated with guidance from the scientific community, policymakers and the public, they claim that AI companies have financial incentives to avoid effective oversight and that current corporate governance structures cannot change this.
What they want is a ‘right to warn’ the public about the risks of AI.
In response, OpenAI said in a statement that the company believes in its scientific approach to addressing risk. The company added that it will continue to engage with governments and others about the technology and has avenues for employees to express their concerns about it.
