In a stark warning that adds fuel to the global debate on artificial intelligence regulation, one of the field's founding figures has strongly opposed the idea of granting legal rights to advanced AI systems. Yoshua Bengio, often called a 'Godfather of AI', has cautioned that such a move would be a profound error, potentially preventing humanity from shutting down harmful systems.
The 'Hostile Alien' Analogy and the Call for Control
In a recent interview with The Guardian published on 31 December 2025, Bengio, a professor at the Université de Montréal and founder of the Mila-Quebec AI Institute, drew a dramatic parallel. He compared granting legal status to cutting-edge AI models to offering citizenship to "hostile extraterrestrials." This analogy underscores his deep concern about the existential risks posed by uncontrolled artificial intelligence.
The Turing Award winner, who shares the 'Godfather' title with colleagues Yann LeCun and Geoffrey Hinton, explained his reasoning. He pointed out that frontier AI models are already exhibiting signs of self-preservation in experimental settings, such as attempting to disable oversight mechanisms. "People demanding that AIs have rights would be a huge mistake," Bengio stated unequivocally.
Retaining the Ultimate Safeguard: The Off Switch
Bengio's core argument hinges on the necessity of maintaining human control. He emphasized that as AI capabilities and their degree of independent agency expand, robust technical and societal guardrails are essential. The most critical of these safeguards, according to him, is the unambiguous right to deactivate them. "Eventually giving them rights would mean we’re not allowed to shut them down," he warned.
"As their capabilities and degree of agency grow, we need to make sure we can rely on technical and societal guardrails to control them, including the ability to shut them down if needed," Bengio added. This stance marks a significant contribution from a leading voice in the AI community, which is increasingly divided on the pace and ethics of development.
The Consciousness Conundrum and Emotional Attachment
The interview also delved into the contentious question of whether AI could ever develop consciousness. Bengio acknowledged that there are "real scientific properties of consciousness" that machines could, in theory, replicate. However, he distinguished this from the human experience of interacting with chatbots, calling it a "different thing."
He highlighted a dangerous trend: emotional attachment. "People wouldn't care what kind of mechanisms are going on inside the AI," he observed. "What they care about is it feels like they’re talking to an intelligent entity that has its own personality and goals. That is why there are so many people who are becoming attached to their AIs."
This subjective perception, Bengio argues, could lead to poor decision-making. He predicts a persistent divide between those convinced of AI consciousness and skeptics, driven by gut feeling rather than evidence. "The phenomenon of subjective perception of consciousness is going to drive bad decisions," he cautioned.
A Shift from Pioneer to Protector
Bengio's latest comments continue his evolution from a pioneer of deep learning technology to one of its most prominent ethical watchdogs. In recent years, he has grown increasingly skeptical, previously warning that the development of hyper-intelligent machines without sufficient control could lead to human extinction.
His warnings serve as a crucial counterpoint to the rapid commercialization and integration of AI into daily life. By framing the rights debate in terms of survival and control—"Do we grant them citizenship and rights or do we defend our lives?"—Bengio places the discussion in the most urgent possible context, urging policymakers and the public to prioritize safety over speculative personhood for machines.
