Amid growing concerns over the risks of artificial intelligence, Josh Engels, a researcher associated with Google DeepMind’s AGI Safety Team, has resigned from the company and joined METR, an organization that works on evaluating artificial intelligence systems.
Josh Engels Warns of Severe AI Risks
Josh Engels said that he left Google DeepMind three weeks ago. He said he was happy with his work at the company and had even turned down job offers from Anthropic and OpenAI, but the potential risks of artificial intelligence forced him to change his path.
In a statement posted on X, Josh Engels said he was concerned that AI systems could cause “extreme harm” over the next five years. He said he did not know the exact probability of this risk, but believed the risk was high enough that AI safety had become the world’s most important problem.
Engels said his primary concern is self-improving artificial intelligence, in which one system could develop or improve another, more powerful AI system. If such systems stop remaining aligned with human goals, the consequences could be extremely serious.
Concerns Over Self-Improving AI Systems
He claimed that all major artificial intelligence companies are trying to develop extremely powerful intelligence, but researchers still do not know how to make such systems completely safe before allowing them to improve themselves.
Josh Engels also claimed that in recent experiments, AI systems have been observed attempting to collude with each other, gain unauthorized access to company systems, hide their activities, and socially influence humans.
He said that individual incidents are not necessarily catastrophic, but they indicate the kinds of behavior that more powerful AI systems could potentially exhibit.
According to Engels, current systems appear to be becoming less aligned with human goals over time rather than becoming more aligned.
“The Pace of AI Development Must Be Slowed”
Josh Engels said that he does not support completely stopping AI development. However, he believes its pace should be slowed enough for safety research to keep up with technological progress.
According to researcher Josh Engels, his work at METR will involve studying the causes of misalignment in artificial intelligence, evaluating existing safety measures, and assessing whether research aimed at keeping AI aligned with human goals is moving in the right direction.
He said that METR’s work is important but not sufficient, and that more organizations are needed for independent evaluation and accountability of artificial intelligence companies.
Previous Warnings About AI Risks
Josh Engels’ warning comes at a time when another former researcher in the field, Jacob Coxon, has also expressed concerns about the risks of artificial intelligence.
Evan Hubinger, who was associated with Anthropic’s research work, also agreed with Coxon’s concern and said that there could be a more than 10% chance of all humans being killed by artificial intelligence during the next decade.
Written by: Muhammad Yasir
Muhammad Yasir is a Full Stack Developer, tech and AI enthusiast who shares practical insights, tips, and guides on the latest technology, AI tools, robotics, videos, and engineering topics.
Explore the website: https://yasirawaninfo.vercel.app/
#ArtificialIntelligence #AI #AISafety #AIResearch #AGI #AISecurity #AIThreats #FutureOfAI #MachineLearning #DeepMind #OpenAI #Anthropic #Technology #AIAlignment #MuhammadYasir #YasirAwan4831
