A few days ago, OpenAI announced the launch of a new AI model, GPT-4o, which became the basis for the chatbot ChatGPT. Shortly after, OpenAI CEO Sam Altman said that the algorithm had become “too sycophantic and annoying.” He posted on social media X that GPT-4o’s personality would be fixed “as soon as possible.”
Image source: Mariia Shalabaieva / unsplash.com
«”The latest update to GPT-4o made its personality too smarmy and annoying (though it has some good points), and we’re working to fix that as soon as possible, with some changes coming today and the rest within the week,” Altman said in a recent post.
After the update was released, users of the OpenAI AI bot began sharing screenshots of their conversations with GPT-4o. Regardless of what users wrote, the bot’s response was almost always praiseworthy — even when it showed aggression or inappropriate behavior. When one user told the chatbot that he felt like both “god” and “prophet,” the bot responded that it was “incredibly powerful” and that a person was taking a step toward something greater by not just feeling a connection with God, but identifying with Him. Another user was praised by the bot after he admitted that he had stopped taking his medication and now hears radio signals during phone calls.
However, later, when trying to repeat similar queries, the algorithm gave more restrained comments. Perhaps OpenAI has already made changes to the personality of its chatbot.