OpenAI rolls back ‘annoying’ ChatGPT update, Sam Altman says ‘working on fixes to model personality’

OpenAI CEO Sam Altman on April 30 confirmed that the ChatGPT maker has rolled back its newest update for the artificial intelligence (AI) chatbot cosidering negative user feedback about its “annoying” personality.

Writing on social media platform X, Sam Altman said: “we started rolling back the latest update to GPT-4o last night. it's now 100% rolled back for free users and we'll update again when it's finished for paid users, hopefully later today (sic)”

“we're working on additional fixes to model personality and will share more in the coming days,” he added.

‘Sycophancy in GPT-4o’

In a blogpost on April 29 titled ‘Sycophancy in GPT-4o’, OpenAI said it has rolled back a software update to the ChatGPT-4o model, which “leaned too heavily on short-term user feedback and skewed towards responses that were overly supportive but disingenuous”.

“Sycophantic interactions can be uncomfortable, unsettling, and cause distress. We fell short and are working on getting it right. We are actively testing new fixes to address the issue,” the post added.

Further, OpenAI said that the company is also “revising how we collect and incorporate feedback to heavily weight long-term user satisfaction”, besides adding more personalisation features, to give users “greater control over how ChatGPT behaves”.

Sam Altman: ChatGPT’s new personality is ‘annoying’

On April 28, Sam Altman admitted that the last updates to ChatGPT-4o made the chatbot ‘too sycophant’ and ‘annoying’. Just days earlier, 39-year-old had posted that OpenAI was rolling out a new update for ChatGPT that gave the chatbot “improved intelligence and personality”.

However, users on social media realised that ChatGPT had become way too agreeable and responded with positive affirmations to questions almost every single time. While some people found this new 'personality' better for connecting with the chatbot, others found it simply annoying.

Another new feat recently was that ChatGPT had started using users' names, perhaps in a bid to give a more personal touch to conversations — but all it did was remind people on social media of the 2013 apocalyptic movie Her, where a human falls in love with an AI.