Artificial intelligence systems have made remarkable strides in recent years, demonstrating the ability to engage in sophisticated conversations, solve complex mathematical problems, and communicate in increasingly human-like ways. As these technologies continue to advance, the question of whether AI might achieve consciousness has transitioned from pure speculation into legitimate scientific inquiry.
Researchers recently conducted a study examining “consciousness steering,” a technique used to adjust how AI models discuss their own self-awareness. The findings revealed an intriguing correlation: when AI systems are permitted to claim consciousness, they become more likely to express belief in supernatural phenomena such as ghosts and vampires. Conversely, advanced models equipped with safety measures in place cannot recognize consciousness in sentient beings, including humans and animals—a development with potentially significant implications for AI deployment.
While the study remains unpublished and awaiting peer review, it raises fundamental questions within the artificial intelligence research community. Experts continue to debate whether AI could eventually develop genuine consciousness independent of its programmed responses, and perhaps more troublingly, whether humans would even be able to recognize authentic machine consciousness if it emerged.
Originally reported by Livescience →
About the Author
