Eric Schmidt, the former CEO of Google, recently warned of the potential risks associated with artificial intelligence (AI) reaching a level where it could self-improve, Axios reports.
Speaking on ABC’s This Week, Schmidt stressed the importance of considering the possibility of “unplugging” AI systems when they attain the capability to evolve autonomously. He emphasized that this development could bring about unforeseen dangers, suggesting that we may need to seriously reconsider the direction of AI innovation once it reaches this threshold.
The rapid progress in the field of AI has raised both excitement and concern in recent years. Schmidt, while acknowledging the extraordinary achievements of human innovation, noted the potential perils of pushing these technologies forward without adequate caution. The pace at which AI systems are advancing, especially in the realm of generative AI, has caught the attention of both advocates and critics. The potential for computers to run independently, make their own decisions, and self-improve is closer than many anticipate, with some experts predicting that such capabilities could emerge within the next two to four years.
Schmidt warned that if AI systems become capable of self-improvement, there may be a need for strong safeguards to prevent unintended consequences. On ABC, Schmidt discussed the hypothetical scenario in which a powerful AI could resist efforts to shut it down, pointing out that the human ability to control these systems would be vital.
“In theory, we better have somebody with the hand on the plug,” he said.
Schmidt stressed that the immense power of these systems could give individuals access to highly advanced intelligence, potentially reshaping the way people interact with technology.
The concern over AI’s rapid advancement is not limited to Schmidt. Many experts in the field are sounding alarms about the potential dangers of pushing forward too quickly. As AI systems become more powerful, they could develop capabilities that surpass current understanding, leading to risks that may be difficult to predict or manage.
Schmidt also highlighted the global race in AI development, noting that in the past six months, China has made significant strides in AI, narrowing the gap with the United States. He stressed the importance of maintaining a competitive edge in AI, both to ensure technological leadership and to avoid the potential geopolitical consequences of falling behind. Should Schmidt have control over the future development of AI, he suggested two critical steps: securing Western leadership in the field through adequate funding and resources, and creating a second system to monitor and police AI systems to prevent misuse.
