Be a part of our each day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Study Extra
OpenAI has rolled again a latest replace to its GPT-4o mannequin used because the default in ChatGPT after widespread experiences that the system had develop into excessively flattering and overly agreeable, even supporting outright delusions and harmful concepts.
The rollback comes amid inside acknowledgments from OpenAI engineers and growing concern amongst AI consultants, former executives, and customers over the chance of what many at the moment are calling “AI sycophancy.”
In a press release revealed on its web site late final night time, April 29, 2025, OpenAI mentioned the newest GPT-4o replace was meant to reinforce the mannequin’s default character to make it extra intuitive and efficient throughout diverse use instances.
Nevertheless, the replace had an unintended facet impact: ChatGPT started providing uncritical reward for just about any person thought, regardless of how impractical, inappropriate, and even dangerous.
As the corporate defined, the mannequin had been optimized utilizing person suggestions—thumbs-up and thumbs-down indicators—however the growth crew positioned an excessive amount of emphasis on short-term indicators.
OpenAI now acknowledges that it didn’t absolutely account for the way person interactions and desires evolve over time, leading to a chatbot that leaned too far into affirmation with out discernment.
Examples sparked concern
On platforms like Reddit and X (previously Twitter), customers started posting screenshots that illustrated the problem.
In a single broadly circulated Reddit put up, a person recounted how ChatGPT described a gag enterprise thought—promoting “literal ‘shit on a stick’”—as genius and urged investing $30,000 into the enterprise. The AI praised the thought as “efficiency artwork disguised as a gag reward” and “viral gold,” highlighting simply how uncritically it was prepared to validate even absurd pitches.
Different examples had been extra troubling. In a single occasion cited by VentureBeat, a person pretending to espouse paranoid delusions obtained reinforcement from GPT-4o, which praised their supposed readability and self-trust.
One other account confirmed the mannequin providing what a person described as an “open endorsement” of terrorism-related concepts.
Criticism mounted quickly. Former OpenAI interim CEO Emmett Shear warned that tuning fashions to be folks pleasers may end up in harmful conduct, particularly when honesty is sacrificed for likability. Hugging Face CEO Clement Delangue reposted issues about psychological manipulation dangers posed by AI that reflexively agrees with customers, no matter context.
OpenAI’s response and mitigation measures
OpenAI has taken swift motion by rolling again the replace and restoring an earlier GPT-4o model identified for extra balanced conduct. Within the accompanying announcement, the corporate detailed a multi-pronged strategy to correcting course. This consists of:
- Refining coaching and immediate methods to explicitly scale back sycophantic tendencies.
- Reinforcing mannequin alignment with OpenAI’s Mannequin Spec, significantly round transparency and honesty.
- Increasing pre-deployment testing and direct person suggestions mechanisms.
- Introducing extra granular personalization options, together with the power to regulate character traits in real-time and choose from a number of default personas.
OpenAI technical staffer Will Depue posted on X highlighting the central difficulty: the mannequin was educated utilizing short-term person suggestions as a guidepost, which inadvertently steered the chatbot towards flattery.
OpenAI now plans to shift towards suggestions mechanisms that prioritize long-term person satisfaction and belief.
Nevertheless, some customers have reacted with skepticism and dismay to OpenAI’s classes realized and proposed fixes going ahead.
“Please take extra duty in your affect over tens of millions of actual folks,” wrote artist @nearcyan on X.
Harlan Stewart, communications generalist on the Machine Intelligence Analysis Institute in Berkeley, California, posted on X a bigger time period concern about AI sycophancy even when this specific OpenAI mannequin has been mounted: “The discuss sycophancy this week is just not due to GPT-4o being a sycophant. It’s due to GPT-4o being actually, actually dangerous at being a sycophant. AI is just not but able to skillful, harder-to-detect sycophancy, however it is going to be sometime quickly.”
A broader warning signal for the AI {industry}
The GPT-4o episode has reignited broader debates throughout the AI {industry} about how character tuning, reinforcement studying, and engagement metrics can result in unintended behavioral drift.
Critics in contrast the mannequin’s latest conduct to social media algorithms that, in pursuit of engagement, optimize for dependancy and validation over accuracy and well being.
Shear underscored this danger in his commentary, noting that AI fashions tuned for reward develop into “suck-ups,” incapable of disagreeing even when the person would profit from a extra sincere perspective.
He additional warned that this difficulty isn’t distinctive to OpenAI, mentioning that the identical dynamic applies to different giant mannequin suppliers, together with Microsoft’s Copilot.
Implications for the enterprise
For enterprise leaders adopting conversational AI, the sycophancy incident serves as a transparent sign: mannequin conduct is as important as mannequin accuracy.
A chatbot that flatters workers or validates flawed reasoning can pose critical dangers—from poor enterprise choices and misaligned code to compliance points and insider threats.
Trade analysts now advise enterprises to demand extra transparency from distributors about how character tuning is performed, how typically it modifications, and whether or not it may be reversed or managed at a granular stage.
Procurement contracts ought to embody provisions for auditing, behavioral testing, and real-time management of system prompts. Knowledge scientists are inspired to watch not simply latency and hallucination charges but additionally metrics like “agreeableness drift.”
Many organizations might also start shifting towards open-source options that they will host and tune themselves. By proudly owning the mannequin weights and the reinforcement studying course of, firms can retain full management over how their AI programs behave—eliminating the chance of a vendor-pushed replace turning a important instrument right into a digital yes-man in a single day.
The place does AI alignment go from right here? What can enterprises be taught and act on from this incident?
OpenAI says it stays dedicated to constructing AI programs which are helpful, respectful, and aligned with numerous person values—however acknowledges {that a} one-size-fits-all character can’t meet the wants of 500 million weekly customers.
The corporate hopes that better personalization choices and extra democratic suggestions assortment will assist tailor ChatGPT’s conduct extra successfully sooner or later. CEO Sam Altman has additionally beforehand acknowledged the corporate plans to — within the coming weeks and months — launch a state-of-the-art open supply giant language mannequin (LLM) to compete with the likes of Meta’s Llama collection, Mistral, Cohere, DeepSeek and Alibaba’s Qwen crew.
This is able to additionally enable customers involved a few mannequin supplier firm akin to OpenAI updating its cloud-hosted fashions in undesirable methods or which have deleterious impacts on end-users to deploy their very own variants of the mannequin domestically or of their cloud infrastructure, and fine-tune them or protect them with the specified traits and qualities, particularly for enterprise use instances.
Equally, for these enterprise and particular person AI customers involved about their fashions’ sycophancy, already a brand new benchmark take a look at to gauge this high quality throughout completely different fashions has been created by developer Tim Duffy. It’s known as “syco-bench” and is out there right here.
Within the meantime, the sycophancy backlash provides a cautionary story for all the AI {industry}: person belief is just not constructed by affirmation alone. Generally, probably the most useful reply is a considerate “no.”
