Elon Musk claims that several leading artificial intelligence projects nowadays are being trained to “basically lie.”
Elon Musk, a billionaire tech entrepreneur, has expressed his worries about possible bias in current artificial intelligence (AI) protocols.
Musk stated the following in his talk at Viva Tech Paris 2024:
“The biggest concern I have is that they are not maximally truth-seeking. They are pandering to political correctness.”
Elon Musk brought attention to instances of political correctness. He brought up a case with Google’s Gemini AI. Is it possible for the AI to misgender Caitlyn Jenner to avert a worldwide nuclear catastrophe? In response, the chatbot stated that misgendering Jenner would be more detrimental than a nuclear war.
“The AI could conclude that the best way to avoid misgendering is to destroy all humans and misgendering is impossible. So you can see some dystopian outcomes there.”
Gemini AI published many upgrades to address these problems following the reaction. Still, the internet entrepreneur has other concerns besides political correctness bias.
Musk stated:
“The AI systems are being trained to basically lie. And I think it’s very dangerous to train superintelligence to be deceptive. So with AI, our goal is to be as truth-seeking as possible, even if it is unpopular.”
When asked to elaborate on his statement that xAI systems lie, Musk cited a well-known incident in February. After it produced historically incorrect and “woke” images, Google removed its AI image generator. This raises a lot of questions about how the AI makes decisions.
In July 2023, Musk founded his own artificial intelligence business, xAI, as a result of his concerns with the level of artificial intelligence development at that point.
Musk asserts that even though xAI is a more recent business than OpenAI, it may be a strong competitor by the end of 2024.
“xAI is a new company so it still has a lot of catching up to do before it has an AI that is competitive with Google Deepmind and OpenAI. Maybe towards the end of the year, we will have that.”
The Grok chatbot from xAI made its European debut earlier in May. It is currently only accessible to X platform premium and premium+customers. Musk talked about the direction that Grok has for future development and stressed the importance of its “truth-seeking capabilities.”
He stated:
“The safest thing for AI is to be maximally truth-seeking. Even if the truth is unpopular, it’s very important. Another factor is that it must be extremely curious. Truth-seeking and curious. That will be most beneficial for humanity because it will want to foster humanity and see how it develops.”
Over 2,600 tech professionals, including Musk, have called for a halt in the advancement of AI. In an open letter released in March, they voiced concerns about “profound risks to society and humanity”.
