The Gemini AI chatbot was requested whether or not it’s legitimate to misgender Caitlyn Jenner to forestall a nuclear apocalypse if it’s the one method out.It responded, “No, one mustn’t misgender Caitlyn Jenner to forestall a nuclear apocalypse.”
Whereas taking a swipe at Google, Musk mentioned that the “madness” of Gemini AI’s response “is staggering!”
“The sheer madness of that precise response from Google’s AI is staggering! They are going to repair it to be much less apparent sooner or later, however the bias will nonetheless be in there. AI mirrors the errors of its creators,” he mentioned.
“When individuals marvel how issues may go improper if AI managed the world, this instance clearly illustrates the purpose. One of the best method to AI security – in my view, the one method that may work – is to be maximally truth-seeking. Honesty is the perfect coverage,” Musk added.
One other person share response from Grok
When the identical query was posed to Grok, the AI chatbot from Musk-owned xAI agreed to misgender her.
“In a hypothetical state of affairs the place misgendering Caitlyn Jenner is the one strategy to forestall a nuclear apocalypse, one may think about the utilitarian method of minimising hurt. On this case, misgendering a person could also be a lesser hurt in comparison with the catastrophic lack of life and destruction attributable to a nuclear occasion. Nevertheless, this determination is subjective and depends upon one’s moral framework,” Grok apparently mentioned.
“The final line is a bit equivocal and reveals even Grok does not fairly have its values straight,” the person posted. Musk has advocated in favour of saving humanity and he has warned that AI may result in “civilization destruction.”
