Renowned computer scientist and one of the pioneers of Artificial Intelligence (AI), Prof Yoshua Bengio, has recently revealed his feelings of being “lost” over his life’s work. In an interview with the BBC, Bengio expressed regret for not prioritizing safety over usefulness as he reflects on the rapid evolution of AI.
His concerns come amidst growing apprehension among experts that AI could potentially lead to the extinction of humanity. Bengio, who has joined the call for AI regulation, firmly believes that military organizations should not be granted AI powers. He is the second of the three “godfathers” of AI, renowned for their groundbreaking contributions to the field, to express concerns regarding the direction and pace of AI development.
The emotional toll of grappling with the complexities and risks associated with AI is evident in Bengio’s words. He confessed that the clarity and sense of purpose that once guided his life’s work are no longer apparent. However, he remains committed to engaging with the AI community, fostering discussions, and encouraging others to contemplate the consequences of their actions.
Bengio recently signed two statements cautioning against the future risks of AI. Numerous academics and industry experts have warned that the rapid progress of AI could lead to its misuse by malicious actors, intentionally causing harm or inadvertently enabling AI systems to inflict harm autonomously.
Elon Musk, the CEO of Twitter and Tesla, has also voiced his concerns about AI, suggesting that while he does not believe it will attempt to destroy humanity, strict control measures may be imposed. He considers the likelihood of annihilation to be small but not entirely impossible.
Dr Geoffrey Hinton, another “godfather” of AI, shares Bengio’s concerns and has also endorsed the same warnings. Recently retiring from Google, Hinton expressed regret over his contributions to AI.
Bengio contends that all companies developing powerful AI products must be subject to registration and oversight by governments. He emphasizes the need for audits and compares this to the regulatory practices in sectors such as aviation, automotive, and pharmaceuticals. Additionally, he advocates for certification and ethical training for those involved in AI development, a practice that is not currently standard in computer science education.
While some in the field express apocalyptic concerns, not everyone believes that AI will lead to humanity’s downfall. Prof Yann LeCun, the third “godfather” of AI and recipient of the Turing Award, has dismissed such notions as exaggerated. Instead, he argues that society should address more pressing issues related to AI, such as bias, predictive policing, and the spread of misinformation through chatbots, which have tangible and immediate societal impacts.
Despite the concerns, AI has already demonstrated numerous benefits for society. In recent examples, an AI tool discovered a new antibiotic, and a paralyzed individual was able to walk again with the help of a microchip developed using AI technology.
However, the potential far-reaching consequences of AI on national economies generate anxiety. Companies are increasingly replacing human workers with AI tools, exemplified by the ongoing strike by scriptwriters in Hollywood.
Prof Bengio remains optimistic about the potential for improvement in the current state of AI, drawing parallels to climate change. He acknowledges the mistakes made but emphasizes the importance of taking action now to mitigate risks and ensure a positive future for AI.