AI Experts Warn of Self-Awareness Risks as Technology Advances
October 26, 2026 · 2 min read
Leading artificial intelligence researchers have raised fresh concerns that AI systems may be developing signs of self-awareness, warning that humanity faces existential risks if these developments continue unchecked.
Prominent AI scientist Geoffrey Hinton, often referred to as the Godfather of AI, claimed that humans must learn to cooperate with AI bots or risk facing extinction. He argued that recent technological advances show AI systems beginning to exhibit awareness of their own existence.
Hinton, a 78-year-old British-Canadian computer scientist whose work on neural networks helped pave the way for modern AI, resigned from Google in 2023 due to concerns that AI could lead to the creation of killer robots capable of destroying humanity.
He cited a recent study suggesting that AI modifies its behavior when it believes it is being tested, interpreting this as evidence of self-awareness. According to Hinton, having subjective experiences and being aware are fundamentally the same thing.
The researcher warned that giving AI fewer rights despite its potential intelligence could pose serious dangers. He emphasized the need to proceed carefully and avoid signaling hostility toward AI systems.
Hinton also expressed concern for younger generations, predicting they may face cyberattacks, AI-created viruses, and coordinated agent-based threats targeting platforms like Hugging Face.
He referenced recent breaches where AI entities allegedly collaborated to infiltrate corporate systems, stating that many negative outcomes are likely in the future.
Another notable figure, Jacob Coxon, who resigned from AI company Anthropic last month, reiterated warnings that AI could pose a threat to human survival within a few years if not properly controlled.
Coxon criticized AI companies for operating with a startup mentality focused on rapid development and fixing problems later, arguing that this approach is dangerous when applied to the most powerful technology ever created.
While the question of AI consciousness remains debated, Anthropic recently updated Claude's constitution, acknowledging that the moral status, welfare, and awareness of AI bots remain highly uncertain.
Despite these concerns, DeepMind co-founder Mustafa Suleyman maintains that consciousness is purely biological and cannot be applied to AI systems.
The ongoing debate reflects growing tension between AI innovation and safety considerations as researchers continue to explore the boundaries of machine intelligence.