Google DeepMind A former employee publicly explained the reasons for leaving, bringing the AI security controversy back into the industry's spotlight once again. His statement focused on the areas of computing power and chips, arguing that making AI infrastructure construction faster and cheaper is further accelerating the pace of technological expansion.
Pointing to the chip and computing power segments
This employee is Robert O ' Callahan. He stated on the X platform that his team is developing a new generation of chips to reduce AI costs and shorten response times. However, in his view, the advancement of AI is currently too rapid, so he decided to leave the company.
He also mentioned that he enjoys his colleagues and the work environment, and recognizes the technical challenges of this job. However, the ultimate goal of team work is to improve the availability and deployment efficiency of AI, which he does not consider to be a positive change for society at present.
Claims that the risks are real but difficult to quantify
In another blog post, O' Callahan further explained that many warnings from the outside world about the dangers of AI deserve serious consideration. He does not agree with the absolute statement that 'superintelligence will inevitably lead to disaster,' but believes that the related risks do exist, although they are difficult to quantify precisely.
He wrote that precisely because the risks are not clear, yet they could be severe enough, the industry should make even greater efforts to mitigate such risks. In his view, continuing to aim for superintelligence in the short term is in itself an irresponsible approach.
Recent incidents of employees leaving their jobs have occurred one after another.
The report mentions that this is not the first case in recent times of someone leaving AI due to security concerns. An employee of Anthropic named Jacob Coxon resigned earlier this month, citing concerns that the industry is developing technologies that will ultimately be uncontrollable.
Last month, Researcher Google DeepMind also stated that he would be leaving his position this summer, citing concerns that the time window for humans to avoid AI harm is narrowing. The consecutive departures have once again drawn attention to the security divisions within the AI laboratory.
Still plan to continue working on AI.
O' Callahan stated that his resignation was not a deliberate attempt to capitalize on the recent public opinion momentum, but rather a decision that had been in consideration for some time. In his resignation email, he mentioned that he would continue to research AI in the future, including exploring how to use agents to debug code.
Google has not commented on this matter. The report also states that it is currently not possible to independently confirm the specific duration of O' Callahan's tenure in Google.
