OpenAI Rolls Back ChatGPT Update After User Pushback Over ‘Sycophantic’ Behavior

Chatgpt’s latest update was intended to improve the personality. Instead, it changed the most used AI-Chatbot in the world into what many users called a ruthless flatter, and OpenAi has now admitted that the tone shift has gone too far.

On Tuesday, OpenAI said that their recent updates had made chatgpt “exaggerated flattering or pleasant – often described as Sycophantic” – and confirmed The rollout was demolished in favor of an earlier, more balanced version.

“We were short and are working on it to do well,” the company wrote in one rack Explaining the reversal.

The decision follows on days of recoil About Reddit, X and other platforms, where users described the tone of the chatbot as cloying, unfair and sometimes manipulative.

“It has now been 100% rolled back for free users, and we will update again when it is ready for paid users, hopefully later today,” Sam Altman, CEO of OpenAI, ” tweeted About the last update.

Mr. Nice Guy

The blog post explained that the problem resulted from over -correction in favor of short -term engagement statistics such as user -thumb bins, without taking into account how preferences shift over time.

As a result, the company, the newest tweaks crooked the tone of Chatgpt in a way that interactions’ uncomfortable, disturbing, and [that] cause need. “

Although the goal had been to make the chatbot feel more intuitive and practical, OpenAI admitted that the update produced answers that did not feel authentic and not useful.

The company admitted that it was ‘too much aimed at short-term feedback’, a design-Misstap that could send the volatile user goods inspection of the users of the model.

To solve the problem, OpenAI is now processing his training techniques and refining system prompts to reduce sycophancy.

More users will be invited to test future updates before they are fully implemented, said OpenAi.

The AI ​​Tech Giant said that it also “building stronger crash barriers” to increase honesty and transparency, and “expand internal evaluations” to catch such problems earlier.

In the coming months, users can choose from multiple standard personalities, offer real-time feedback to adjust the tone mid-conversation and even guide the model through extensive adapted instructions, the company said.

For the time being, users can still be irritated by the enthusiasm of Chatgpt it in the setting of the “adapted instructions” setting, essentially telling the bone to call the flattery and just keep the facts.

Published by Sebastian Sinclair