On Tuesday, a British engineer at Anthropic who has also worked for the past three years at OpenAI, the creator of the popular ChatGPT, resigned, saying that both companies “are not acting responsibly” and are “gambling with our lives.”
RELATED:
Major U.S. Artificial Intelligence Platforms Hit by Near-Simultaneous Outages
Jacob Coxon, a software engineer whom media outlets identify as one of the developers of GPT-4o and GPT-4.5, said that the people behind the exponential development of AI “believe it could kill us all by the end of this decade” because “there is no other human activity that poses this level of threat.”
For the 27-year-old engineer, the power of the technology is being underestimated, but AI projects “will soon be superhuman systems that will be able to hack anything, revolutionize every field and gain real power and resources.”
In his opinion, there is an awareness of this reality at both OpenAI and Anthropic, but he sees a difference: The former “may not have deeply internalized the civilizational challenges,” while the latter understand them well but “are trapped in the race to get there first,” even if, in the most optimistic scenario, getting there first means “acting responsibly.”
Coxon sees no other way forward than coordination among the various AI projects and the companies involved in them, but he feels that this is not happening.
He therefore advocates for “costly actions, such as a temporary ban on improving ‘model capabilities,’” and calls on AI researchers to reflect deeply on the issue.
“Do you want to power a superintelligent ‘reinforcement learning’ (RL) system without a rigorous understanding of its mind?” he asks his fellow AI researchers.
“Are you going to put your head down because ‘well, it’s already happening,’ or take this moment to advocate for different conditions?” Coxon concluded.
Evan Hubinger, an Anthropic researcher, responded to Coxon’s comments and qualified his argument: “Jacob is right here; we genuinely believe that AI could kill all humans. Personally, I think it’s more than 10% within the next decade.”
Despite this, Hubinger considered the risk posed by current models to be low, although what is concerning is “the superintelligence emerging from recursive self-improvement… which is happening faster than we thought.”
Coxon is not the only researcher to have withdrawn from the development of AI models in recent years after warning about the dangers of this technological development.
This is the case with Geoffrey Hinton, a Nobel Prize in Physics winner, who resigned from his position at Google in 2023 so he could criticize the risks of AI without corporate constraints.
The open letter that former OpenAI employees issued in 2024 should also be mentioned. In the letter, they expressed concern about the rapid development of AI and the absence of laws protecting whistleblowers.
teleSUR/ JF
Source: EFE
Source: teleSUR
