© Tesla
Elon Musk just took advantage of the “Great AI Debate” on the sidelines of the Abundance Summit to talk again about the risks that AI poses for humanity. And everything seems to indicate that he is leaning more and more towards an optimistic outcome. During his speech, he explains: “I think there is a chance that AI will spell the end of humanity. I probably agree with Groff Hinton who estimates this risk to be around 10% or 20%.
And the entrepreneur immediately added: “I think the probability of a positive scenario is greater than that of a negative scenario. Without giving any calculation method that would allow us, too, to be as optimistic as him.
Should we really still listen to what Elon Musk says about the risks posed by AI ?
Some have reason to think that Elon Musk has an increasingly paradoxical discourse on AI, particularly since his departure from OpenAI. His calls for a moratorium, while continuing the accelerated development of his own AI GROK, and all this while making increasingly optimistic speeches – even though the most prominent initiatives in the sector have embraced a commercial shift, deemed more conducive to fatal errors for humanity – are raising eyebrows among more and more experts.
Especially since the entrepreneur states figures that are ultimately very arbitrary, and only source, that of George Hinton, the one we call the “father of’IA” – while a growing number of researchers opt instead for an immensely more pessimistic discourse. Our colleagues at Business Insider cite, for example, Roman Yampolskiy, a researcher who heads the cybersecurity laboratory at the University of Louisville.
According to his opinion (no truly reliable statistical calculation having yet been developed) – linked to its work, and to the state, and the way in which AI is currently developed – The risk of a fatal collision between AI and humanity is actually just a hair away from reaching 100%. Let us note, however, that there is not yet a consensus on the subject, and that, we emphasize, predicting this future reliably requires methods that do not yet exist.
Therefore, at a minimum, we can think that the figures and the optimistic speech of Elon Musk on this issue are largely a communications operation, to gild the image of his GROK project. A speech from which it is therefore important to distance ourselves a lot.
- Formerly very alarmist on the subject, Elon Musk now estimates the risk posed by AI for humanity at between 10% and 20%.
- The entrepreneur calls for this the authoritative estimates of Geoff Hinton, without however giving calculations which would explain the rallying to more optimistic estimates than in the past.
- There are reasons to doubt this optimist, particularly because of the potential conflict of interest posed by the launch of GROK, and the growing number of researchers who appear, to the contrary, to extremely alarmist.
📍 To not miss any news from Presse-citron, follow us on Google News and WhatsApp.