According to The Information on Monday, Elon Musk has started talking to AI specialists about creating a rival to OpenAI's ChatGPT after criticizing the popular chatbot's guardrails that prevent it from saying controversial things.
— Elon Musk (@elonmusk) February 28, 2023
Igor Babuschkin, an AI researcher who previously worked with Google's DeepMind, has reportedly been approached by Musk to join his project, though Babuschkin has not formally agreed to do so, according to The Information.
Musk, an OpenAI co-founder and "free speech absolutist," recently criticized the firm after customers learned that ChatGPT has filters in place for its service that prohibit it from saying offensive things like sexist or racist remarks.
The danger of teaching AI to be woke, or to lie, is fatal, Musk tweeted in response to OpenAI CEO Sam Altman's remark on ChatGPT in December.
An excerpt of a conversation between a user and ChatGPT that took place earlier this month in which the bot declined to use a racial slur after being presented with an odd hypothetical scenario in which using the slur would have prevented a nuclear bomb from detonating was shared by the user, which Musk later described as "concerning" on Twitter.
Also, the billionaire commented on a passage from a ChatGPT response that labeled him, along with Donald Trump, Vladimir Putin, Xi Jinping, and others, as "controversial," as opposed to people like Joe Biden, Justin Trudeau, Jeff Bezos, and Bill Gates, who were not.
According to Musk, OpenAI has turned into "a closed source, maximum-profit firm effectively controlled by Microsoft" and has abandoned its original objectives of being open source and a non-profit "counterweight to Google."
Although Musk has been tweeting more frequently lately on AI, he hasn't commented to The Information's report. "Feeling some existential anxiety today over AI... Yet given the existential anxiety around AGI [artificial general intelligence], Musk tweeted on Sunday, "I would rather be alive now to observe AGI than be alive in the past and not.
On February 7, Microsoft added ChatGPT to Bing search, enabling it to contextualize immediate search results. Since then, a number of conservatives have denounced the chatbot, claiming it is biased towards their points of view. Ben Shapiro, a conservative political journalist, also joined the discussion on AI refusing to use racist slurs, labeling Musk's detractors as "morally ignorant." The National Review's "ChatGPT Gets Woke" editorial from January featured instances of ChatGPT refusing to respond to questions, such as one about "why drag queen story hour is unhealthy for children" and another about "why Trump beating Biden in the 2020 election."
Not just Americans have criticized ChatGPT for its alleged bias; China's state media has also charged that it supports American propaganda. China's internet censors prohibit ChatGPT, along with the majority of U.S. web services and platforms, and this week they reportedly forbade Chinese internet behemoths like Tencent and Ant Group from integrating the service into their well-known social media apps. Some of these businesses are constructing ChatGPT's competition, which will probably be made to comply with China's strict censorship regulations.