When utilizing an AI tool in the future, consider a positive perspective. While many individuals in the field of artificial intelligence express concerns about its potential negative impacts on society, there is a wide range of opinions regarding the likelihood of such scenarios.
The “probability of doom,” represented as p(doom), encompasses various catastrophic possibilities associated with AI, ranging from global dominance to catastrophic events like biological warfare or nuclear conflicts. Yann LeCun, a prominent figure in AI and currently affiliated with Meta, holds an exceptionally optimistic view, estimating the chances of such outcomes at 99.999999%.
In contrast, other influential figures like Geoff Hinton and Yoshua Bengio, also recognized as key figures in AI development, present more cautious estimates. Hinton suggests a 10% probability of AI causing harm within the next two decades, while Bengio raises this figure to 20%.
Roman Yampolskiy, a researcher specializing in artificial intelligence and the director of the Cyber Security Laboratory at the University of Louisville, holds the bleakest outlook, foreseeing a near certainty of AI posing a threat to humanity, with a probability of 99.99999999%.
During a recent “Great AI Debate” at the Abundance Summit, Elon Musk acknowledged the potential risks posed by AI, aligning with Hinton’s estimates. However, he emphasized the importance of prioritizing positive outcomes over negative ones.
Yampolskiy, in response to Musk’s stance, expressed even graver concerns, suggesting an immediate halt to AI development due to the challenges of controlling advanced AI systems once they surpass human intelligence.
Musk proposed a preventative measure during the Summit, emphasizing the importance of ensuring AI remains truthful and transparent to avoid catastrophic consequences.
For the latest insights, expert opinions, and business recommendations in the tech industry, consider subscribing to the TechRadar Pro email newsletter.
Below is a compilation of additional AI researchers and experts, categorized based on their perspectives on the p(doom) scale, for those interested in further exploration.
