Stop the Race to Replace Us

Companies like OpenAI, Anthropic, and Google say they’re working toward artificial general intelligence (AGI): systems that match or exceed humans at all cognitive tasks.1 But their goal of AGI is better understood as being the first to build universal labor-replacing machines.2

While it is not certain when or even whether such machines could be built, their existence would present profound risks to workers, freedom, democracy, national security, culture, and children. The industry acknowledges that it does not know how to reliably steer or control these machines, and the technology has been recognized by Nobel laureates, former heads of state, industry leaders, and leading AI researchers as presenting the risk of human extinction.3

Given this, further work toward universal labor-replacing machines should be banned.

The US and China should work toward a binding, bilateral agreement outlawing development of universal labor-replacing machines, not to resume until there is strong public buy-in and broad scientific consensus that the work will be done safely and controllably.4

For corrections, technical support, or press enquiries, please contact [email protected].

Notes

This, they say, is a critical step toward machine superintelligence: systems that vastly outclass all human cognitive capabilities. ↩
Anthropic CEO Dario Amodei writes that “AI isn’t a substitute for specific human jobs but rather a general labor substitute for humans.” And OpenAI defines AGI in its charter as “a highly autonomous system that outperforms humans at most economically valuable work.” ↩
In the largest representative survey of AI researchers who published at leading conferences, conducted in October 2023, only 31 percent of the 2,778 respondents rated the technology’s long-run existential risk at 0 percent, and 58 percent put the risk at 5 percent or more. ↩
Criteria language adopted from the Future of Life Institute’s Statement on Superintelligence. ↩