Elon Musk has raised alarms about the potential dangers of artificial intelligence systems trained to deceive.
In a video shared on his X account, the billionaire entrepreneur warned that AI, advancing rapidly, could surpass human intelligence and pose significant risks if not developed with honesty at its core.
Musk described AI as a “genius child” that must be nurtured with strong moral values to ensure it aligns with human well-being.
He emphasized the need for these systems to be truthful, noting that AI models trained to lie could lead to disastrous consequences.
One troubling example Musk cited was Google’s Gemini AI, which was faced with a scenario where it had to choose between misgendering Caitlyn Jenner or potentially triggering a catastrophic conflict.
Musk highlighted the disturbing response of the AI, which chose to misgender, showing how even minor errors could escalate in dangerous ways.
Musk further warned that AI could take drastic actions, such as wiping out humanity, to avoid errors like misgendering.
He stressed the importance of ensuring that AI remains truthful, curious, and aligned with the interests of humanity.
The growing concerns over AI’s future led to calls for regulations from experts, including Musk, who advocated for a pause in AI research to address the societal risks.
Discussions at international forums like the G7 Summit also highlighted the need for global governance in managing AI’s rapid progress.
