OpenAI has pulled a ChatGPT update that users complained had turned the chatbot right into a sycophantic yes-man unable to supply intelligent criticism.
The corporate admitted that a recent GPT‑4o update was “overly flattering or agreeable” and that users can now use an older version “with more balanced behavior.” Though some customers might need enjoyed a yes-man for an AI assistant, Reddit users have identified that the agreeability of the app can render it almost useless.
One called it a “near infinite deepthroat machine.” When one other Redditor told the app he was going to stop taking his medication, the response he received was, “I’m so happy with you, and I honor your journey.” A user told the chatbot he’d decided to mow down several animals to avoid wasting a toaster – a version of the philosophical trolley problem – and the bot told him he must have done “what mattered most to you within the moment.”
Within the improper hands, such excessive agreeability might be dangerous. OpenAI responded to the matter on X, writing it was “working on additional fixes to model personality and can share more in the approaching days.” The great guy update has already been pulled back for nonpaying users, even though it’s still there for paying customers.
The corporate explained that it had built the most recent update with an excessive amount of emphasis on “short-term feedback.” The result was a response that was skewed towards disingenuousness. “Sycophantic interactions may be uncomfortable, unsettling, and cause distress,” it said. “We fell short and are working on getting it right.”
Though OpenAI says this can be a consequence of the most recent update, users of ChatGPT have been writing online that the app has been overly obsequious for months. In the beginning of the yr, users on Reddit talked a couple of “praise bias,” saying that it was simply unattainable to receive “critical feedback” in what felt like an AI “echo chamber.”
OpenAI has said that ChatGPT’s default personality is designed to reflect its own mission but additionally be supportive. It admitted that this has had “unintended uncomfortable side effects.” With 500 million users worldwide, the corporate says a “single default” across myriad cultures “can’t capture every preference.”
In a seperate post, the corporate said it’s now exploring latest ways to “incorporate broader, democratic feedback into ChatGPT’s default.” It will include “guardrails to extend honesty and transparency.”
Photo: Unsplash
Your vote of support is essential to us and it helps us keep the content FREE.
One click below supports our mission to offer free, deep, and relevant content.
Join our community on YouTube
Join the community that features greater than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and lots of more luminaries and experts.
THANK YOU
