Google DeepMind researcher resigns due to concerns about pace of AI development

(San Francisco = Yonhap News) Correspondent Kwon Young-jeon = Following Antropic, an AI researcher at Google also resigned, raising concerns about the pace of artificial intelligence (AI) development.

According to the U.S. AI industry on the 26th (local time), Robert O'Callaghan, a member of Google DeepMind, Google's AI organization, confessed on the 24th through social networking service (SNS)

In his resignation letter published on his personal blog, he, who lives in New Zealand, explained the reason for quitting the company by saying, "Our team's ultimate goal is to make AI much cheaper and reduce latency," and "I don't think this is good for people at the moment."

“With millions of people contributing to accelerating AI progress, the impact of my stepping out will be very minimal,” he said.

He said he tried to ignore AI developments, but decided to resign because he thought this was not an attitude of following Jesus, revealing that his religious faith was also a factor in his decision to resign.

He confessed that although he was not directly in charge of AI functions and was working on improving tools for hardware chip design, God made him face the reality that the tools would eventually accelerate AI chip design.

He emphasized, “I am not sure that the probability of destruction caused by super artificial intelligence (ASI, AI with capabilities superior to humans in most areas) is 100%,” but added, “I believe the risk is real but uncertain, but that fact alone is very concerning.”

He also pointed out that many of the phenomena predicted by pessimists, such as 'compensation hacking' and 'alignment mismatch', have actually occurred.

He added that most of the people he spoke to at DeepMind were seriously concerned about AI.

However, he explained that his decision to leave Google at this point is due to his personal travel schedule and is not related to the recently raised 'AI speed control theory' or other resignations in the AI ​​industry.

Previously, Jacob Coxon, a researcher who worked at Antropic after working at Open AI, made headlines by announcing his resignation, warning that AI could destroy humanity within 10 years.

After Coxon's resignation, Antropic CEO Dario Amodei argued for AI speed control, and since then, the AI ​​industry has been debating it, divided into those who support it and those who oppose it.

The British daily Financial Times (FT) reported on the 22nd that the number of AI-related employees and researchers resigning or taking sick leave due to low morale and ‘burnout’ is increasing.