Elon Musk Highlights ChatGPT's Hallucinations, But Not Everyone Is Convinced

Elon Musk has brought attention to the issue of chatGPT’s hallucinations, sparking a heated debate among netizens who hold different viewpoints on the matter. 

What Happened: Musk, on Tuesday, engaged in a Twitter discussion regarding ChatGPT, an AI language model developed by OpenAI – the company he co-founded in 2015 and then left in 2018. 

It started when Greg Brockman, another OpenAI co-founder, tweeted about the benefits of asking “dumb” questions as a way to learn. Musk then questioned when AI would surpass humans in answering questions.

See Also: Elon Musk Has A One-Word Response To ChatGPT’ Jailbreak’s’ Philosophy On God

In the same thread, a Twitter user said that a set of 20 diverse questions could be answered better by ChatGPT than any single human, praising the model’s breadth of knowledge. 

However, the tech billionaire countered by expressing his skepticism, pointing out that ChatGPT’s answers may sound correct but can be confidently wrong.

Many tweeps joined the conversation. One user highlighted the significant augmentation provided by ChatGPT, stating that individuals greatly benefit from its wide-ranging knowledge across various domains where their expertise may fall short. 

Another user acknowledged the potential of guiding ChatGPT toward accurate answers. Musk’s parody account also chimed in with a witty remark, stating, “And confidently biased!” 

This comment added a touch of humor to the conversation but also steered the conversation toward the concerns around inherent biases in AI systems. 

Why It’s Important: OpenAI’s ChatGPT previously faced allegations of being “woke,” while Microsoft Corporation’s MSFT Bing AI, which utilizes the same OpenAI technology as ChatGPT, came under scrutiny for deviating from expected behavior. 

Additionally, Alphabet Inc.’s GOOG GOOGL Google Bard received ridicule for providing incorrect information upon its launch, triggering a subsequent debate regarding the accuracy of the chatbot.

Google CEO Sundar Pichai has also acknowledged that AI generally struggles with “hallucination problems” with no clear explanation. 

Check out more of Benzinga’s Consumer Tech coverage by following this link.

Read Next: Elon Musk Credits Himself For ChatGPT-Parent’s Success: ‘I Am The Reason OpenAI Exists’