Geoffrey Hinton, widely regarded as the “Godfather of AI” because of his groundbreaking work on neural networks, has become a more outspoken critic of the AI industry that he helped create.
He recently decided to leave his position at Google, where he had been a senior researcher since 2013. Geoffrey was alarmed by the rapid and unregulated development of generative AIs, such as ChatGPT and Bing Chat, that can produce realistic and persuasive texts, images, and sounds. He feared that these AIs could pose serious ethical and social risks if not used responsibly.
Hinton shared his doubts more widely at the Collision conference in Toronto, which happened this week. He challenged the optimistic views of many companies that AI can handle various tasks with his own cautions. He is not certain that friendly AI will overcome harmful AI, and he thinks that using AI ethically may demand a high cost.
Hinton is of the opinion that large language models (AI systems that create human-like text, such as OpenAI’s GPT-4) could enhance productivity enormously, but he fears that the powerful might use this to benefit themselves, expanding an already huge gap in wealth. Hinton said that it would “make the rich richer and the poor poorer”.
Hinton also repeated his well-known opinion that AI could be a danger to the survival of humanity. He said that there is no assurance that humans will stay in charge if artificial intelligence becomes more intelligent than them. He said that it would be a problem if AI decided that it needed to take over to achieve its objectives. These risks are not just imaginary scenarios, but real possibilities that need to be considered seriously, Hinton said. He said that he was afraid that society would only try to stop killer robots after they saw how terrible they were.
Hinton also expressed his concern about the issues of bias and discrimination that AI can cause. He said that if the data used to train AI is not representative or balanced, it can lead to unfair outcomes. He also said that algorithms can create bubbles that amplify false or harmful information and affect people’s mental health.
Hinton also said that he was worried about AI spreading misinformation beyond those bubbles. He said that he did not know if it was feasible to identify and flag every false claim, even though it was “important to mark everything fake as fake.”
The post ‘Godfather of AI’ Geoferry Hinton doesn’t believe good AI will overcome bad AI appeared first on Techlusive.
from Techlusive https://ift.tt/GsCNFYB
via
IFTTT
Comments
Post a Comment
add