The potential for human extinction due to artificial intelligence is a topic of increasing concern among researchers. In a recent interview, Geoffrey Irving, a former employee of OpenAI and Google DeepMind, suggested that the odds of human extinction could be as high as 50%. This alarming perspective is part of a series of interviews conducted by Palisade Research, a nonprofit organization focused on examining the capabilities and motivations of AIs. The interviews feature insights from various individuals associated with major AI companies, including OpenAI, Google, and Anthropic, and can be accessed at frominside.ai.
Irving’s views are echoed by others in the field. Neel Nanda, a research scientist at Google DeepMind, expressed his concern, stating there is “at least a 10 percent chance” that AI could lead to human extinction, labeling this risk as “ridiculously high.” Both Nanda and Irving have been referenced in a recent report emphasizing the challenges of defining AI safety and the complexities surrounding actionable solutions in this area.
Former OpenAI researcher Daniel Kokotajlo added to the discourse, describing superintelligent AI as “god-like powerful.” He warned that without knowledge of how to regulate such entities, controlling them could be nearly impossible. Kokotajlo called for serious attention to this issue, characterizing the potential risks as significant and urgent.
The interviews available include the full content as well as curated segments that address a range of topics, such as whether the dangers of AI are being overstated and the motivations behind working on such perilous technology. Notably, Mary Phuong from Google remarked on the need for skepticism regarding her statements because of her employment by the lab. Nanda further explained his commitment to the field, asserting that he would resign if he did not believe his work mitigates existential risks, yet acknowledged that companies would proceed with AI development regardless of individual departures.
Editor’s Take
This dialogue among leading AI researchers sheds light on the growing tension between innovation and safety. As AI technology evolves, understanding the risks associated with it becomes critical for developers, businesses, and society at large. The stark warnings from experts may prompt stakeholders to take a more proactive approach towards ethics and safety in AI development, ensuring that advancements do not come at the expense of human safety.
Source: www.theverge.com
