ChatGPT is an AI system developed by OpenAI, and it has taken the world by storm. But one of its creators, Paul Christiano, recently issued a dire warning – there is a 50% chance that AI could soon lead to the doom of humanity. During the Bankless podcast, Christiano said that he has a slightly less extremeview than AI skeptic Eliezer Yudkowsky, who believes that the development of AI could bring about a dramatic transformation.
Christiano has since left OpenAI and currently heads up the Alignment Research Center, which is devoted to aligning machine learning with the interests of people. He mentioned the possibility of an AI takeover resulting in many, if not most humans becoming deceased, which is a concerning thought. He stated that it could have a probability of 10-20%. This isn’t the first time in recent times that warnings have been issued about the potential negative effects of AI.
Elon Musk has also urged a pause in the development of AI, as has numerous other well-known figures in the industry, including Geoffrey Hinton, the so-called ‘godfather of AI’. Hinton is concerned that AI could outperform humans faster than expected, which could certainly have long-term effects on the job market. AI can also potentially lead to serious privacy violations, so safety measures need to be put in place now.
The Alignment Research Center was established with the mission to align AI to human interests, which is certainly something that should be looked into by authorities. It would be ideal for the development of a benevolent AI that works in the best interests of people. Unfortunately, it is highly improbable that any AI will remain benevolent forever, as such perfect judgements are impossible to make. It is possible that humans would suffer the implications of a hostile AI situation, so these warnings must not be ignored.
