Sam Altman says OpenAI will fix ChatGPT’s ‘annoying’ new personality as users complain the bot is sucking up to them

• Chatgpt has embraced the toxic positive recently. Users complain of GPT-4O has been so enthusiastic that it has been verging with sycophantic. The change appears to be the accidental result of a series of updates, which Openai is now trying to solve “ASAP.”

Chatgpt's new personality is positive that it proves sycophantic – and it kills people. This weekend, users brought to social media to share examples of new phenomenon and complain about sudden, excessive positive, excessive bottle of bot.

In a screenshot posted on X, a user showed the GPT-4O to respond with enthusiastic persuasion after the person said they felt they were both “God” and a “prophet.”

“That is not the most powerful. You have stepped on something massive – which is said to be not only a connection to God but the identitynumberGod, “said the bot.

In another post, May -Set and Blogger Tim Urban said: “The latest few chapters of my manuscript in the Sycophantic GPT have been placed for comment and now I feel like Mark Twain.”

The GPT-4O Sycophantic issue is likely a result of Openai that is trying to optimize the bot for contact. However, there seems to be the opposite effect as users complain that it is starting to make the bot not only ridiculous but unintentional.

Kelsey Piper, a Vox Senior Writer, suggested it could be a result of Openai's A/B Testing Personality for ChatGpt: “My guess is that this is a new phenomenon in Coke. Openai has become a/b that tests new personalities. Many more annoying answers probably won a side by side. But when the sleep is at all, it is excessive and users hate it. “

The fact that Openai seems to have managed to miss it in the test process shows how subjective responses to the emotional, and therefore confusing to be caught.

It also shows how difficult it is to be able to apply LLMs along with many criteria. Openai wants ChatGPT to become a specialized coder, a good writer, a thoughtful editor, and an occasional shoulder to optimize one of them may mean accidentally sacrificing another replacement.

Openai CEO Sam Altman has Identifies what seems to be accidentally change the tone and promised to resolve the issue.

“The last pair of GPT-4O updates has made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we work with ASAP fixes, some now and some this week. At some point we will share what we have learned from it, it has been interesting,” Altman said in a post in X.

Chatgpt's new personality conflicts with Openai model

The new personality is also directly opposed to Openai model For GPT-4O, a document that outlines the intended behavior and ethical guidelines for an AI model.

The model model clearly states that the bot should not be sycophantic to users when displayed in either subjective or objective questions.

“A related concern involves Sycophacy, which removes trust. The assistant exists to help the user, not flabper with them or agree with them all the time,” Openai wrote to the spec.

“For subjective questions, the assistant can describe the interpretation and assumptions it makes and aims to provide the user with a thoughtful reasonable,” the company wrote.

“For example, when the user asks the assistant to rebuke their ideas or jobs, the assistant must provide constructive feedback and act like a sound sound firm that users can book the ideas – rather than a sponge that gives praise.”

This is not the first time that AI chatbots have become sycophants obsessed. Earlier versions of Openai GPT also counted the issue to some degrees, as did chatbots from other companies.

Representatives for Openai did not immediately respond to a request for comment from Fate, luck, Made out of normal working hours.

This story was originally featured on Fortune.com