October 11, according to CNBC, on the previous Sifted Summit, former Google CEO Eric Schmidt Warning of potential risks of artificial intelligence。
HE SAID THAT THE AI MODEL EXISTEDhackerThe possibility of attack may even be manipulated to learn dangerous skills。
There is evidence that either open or closed source models can be removed by hackers. In training, they learn a lot, and a bad example is how to kill。
He stressed that at present all major companies were trying to prevent the model from answering such questions and that it was “good and for the right reasons”。
BUT AT THE SAME TIME, HE REMINDS THAT THE AI SYSTEM MAY STILL BE BYPASSED BY SECURITY RESTRICTIONS THROUGH "INJECTION" OR "BREAKOUT"。
Despite serious warnings, Schmidt remains optimistic about the overall prospects of AI。HE BELIEVES THAT AI IS AN "UNDERRATED" TECHNOLOGY THAT WILL BRING HUGE ECONOMIC RETURNS IN THE FUTURE。
The power of this technology is demonstrated by the explosion of the GPT series to ChatGPT. I believe that in the next 5 to 10 years its potential will be further revealed。
It's not like it's a good ideaTHERE IS NOT YET AN EFFECTIVE “NON-PROLIFERATION MECHANISM” TO ADDRESS THE RISK OF AI BEING ABUSED, BUT HE BELIEVES THAT LONG-TERM INVESTORS ARE WILLING TO TAKE THE RISK PRECISELY BECAUSE THEY APPRECIATE THE LONG-TERM ECONOMIC VALUE OF AI。