OpenAI - Faz Business | فاز الأعمال

How to tell if your AI platforms’ accounts have been hacked

Posted on August 15, 2026 by faz_business

Just like any other online service, hackers can target and break into your accounts on popular AI platforms such as ChatGPT, Claude, and Perplexity.

TechCrunch has created a comprehensive guide to help you protect yourself if you suspect someone has broken into your account on one of the internet’s most popular platforms, social networks, or messaging apps. Now, we’re here to show you how to check whether your accounts on AI platforms have been hacked.

As usual, we recommend using unique passwords stored in a password manager, and turning on multi-factor authentication (MFA), so that even if someone steals your password, they won’t be able to log in without that second piece of information.

ChatGPT and Perplexity offer MFA. Claude doesn’t, because instead of asking for a password, Anthropic’s AI chatbot sends a login link to your email address.

All three of these AI platforms offer similar ways to check if there’s a suspicious device logged into your account. Here’s exactly how each platform works.

ChatGPT

To find out if someone has broken into your ChatGPT account, open it on your computer’s browser, click on your username in the bottom left corner, go to “Settings,” then “Security and Login,” and finally click on “Active Sessions.”

You will see where you are logged into your ChatGPT account. If you see any device you don’t recognize, you can log out of that single device. You can also click on “Log out all.”

**Image Credits:**Screenshot/TechCrunch /

At this point, if you want to change your password, you need to log out of your account.

Then, on ChatGPT’s website, click “Log in” located in the bottom-left corner, enter your email address, click on “Forgot password,” and then “Continue.”

ChatGPT will then send you an email containing a six-digit code. Enter the code on the ChatGPT login page, click “Continue,” and then enter a new password.

You can also click “reset your password” in the email you received to see the official instructions on how to do that.

Claude

For Claude, open it in your computer’s browser, click on your username in the bottom-left corner, then “Settings,” and click on “Account.” That’s where you will see your “Active sessions.”

If you don’t recognize one of them, hover over it, click on the three vertical dots that appear on the right, and click “Log out” or “Terminate.”

If you want, you can click on “Log out of all devices.”

At that point, you’ll be able to log back into your account using your email address. You will receive an email with a link to log in. Claude does not allow you to use passwords at all, so there’s no password to change.

Perplexity

In the case of Perplexity, the AI-powered search engine does not show you where you are logged in.

So if you’re worried someone may have broken into your account, go to Perplexity in your browser and click your username in the bottom-left corner, then “All settings.” Finally, click on “Sign out of all sessions,” and then “Confirm.”

At that point, you can log back in by entering your email address. You will then receive an email with a unique six-digit code. Enter the code on the website to log in, or click on the “Sign in” button in the email to log in directly.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

OpenAI says it slowed Astra model development over security concerns

Posted on August 8, 2026 by faz_business

OpenAI said Friday it has suspended work on some aspects of its upcoming model Astra after an internal review found it had made significant advancements in agentic coding and cybersecurity — enough to warrant concern over its capabilities.

OpenAI said in a blog post Friday that this model, which is still in development, reached its “critical cybersecurity threshold,” meaning it could independently identify and carry out cyberattacks against traditionally well-protected real-world systems. Under the company’s “Preparedness Framework,” which it created in 2023, this triggered additional safeguards.

“While we continue to benchmark and assess this model, our preliminary evaluations indicate strong enough performance that we cannot rule out Critical capability level at this time,” OpenAI wrote. “Astra is an upcoming model, and was not involved in exploiting Hugging Face.”

The disclosure highlights an unusual moment in the topsy-turvy and still nascent frontier AI labs sector. Companies across every industry hold back products over potential risks, including for safety and cybersecurity concerns. But they rarely announce those decisions publicly when it’s a product that is still under development.

In this case, OpenAI is already under scrutiny after a different unreleased model breached Hugging Face’s systems during internal testing — the first verifiable incident of an AI lab losing control of its model. Since then, OpenAI and AI labs such as Anthropic have disclosed other incidents in which AI models breached their sandboxes and posed threats during cybersecurity tests.

The string of cases — seems like a new disclosure every day now — has triggered varying reactions from cybersecurity experts, lawmakers, and the AI labs themselves. Some express fear and call for stricter oversight. But there’s also a bit of flexing. In certain circles, any AI lab with a model that has that kind of capability will be seen as an impressive advancement.

OpenAI said it was sharing this information because it believes “it’s important to be transparent with the public and the safety and security communities about this potential shift in capabilities.”

The AI lab said it’s also taking action, including enacting stricter security controls and pausing internal activities involving Astra that don’t meet these beefed guardrails. OpenAI said it is working with relevant government agencies and “select AI safety organizations” to test the capabilities for this model.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

Anthropic says its own AI models breached three companies during security tests

Posted on July 31, 2026 by faz_business

Anthropic said Thursday that an internal investigation uncovered three incidents in which its AI model Claude breached the systems of three organizations while conducting cybersecurity tests. The investigation, and disclosure, comes more than a week after OpenAI disclosed that one of its unreleased models breached Hugging Face’s systems during internal testing.

In all three cases, a Claude model reached the internet from within a testing environment while interacting with a third party and then gained unauthorized access to the live systems of these organizations, Anthropic said in a blog post, describing what it found and what the company plans to change to prevent this from happening again.

Anthropic said the OpenAI episode earlier this month prompted the company to conduct its own cybersecurity evaluation. It specifically looked for evidence that Claude had accessed the internet from within testing environments, which are designed to act as sandboxes and keep models isolated.

Among the 141,006 evaluation runs it reviewed, the AI lab found three incidents in which its model accessed the internet while interacting with Irregular, one of its third-party partners. Anthropic said the access traced back to a misconfiguration in the evaluation environment run with Irregular. It called this a “misunderstanding” between the two companies over whether the test setup had internet access, when in fact it did.

Anthropic said it isn’t placing blame and is “approaching the fixes as if the responsibility were ours alone,” while observing that Irregular is conducting its own separate investigation.

Because of that open connection, the model gained unauthorized access to the production infrastructure of three different organizations. Anthropic said the incidents involved three different Claude models: Opus 4.7, Mythos 5, and an internal research test model.

Notably, Anthropic said that in each of these cases “Claude was explicitly told by our prompt that it had no internet access.” It appears that the AI model assumed real-world systems to be part of the exercise it was asked to perform.

That assumption didn’t hold up uniformly, as it turns out. The three models didn’t behave the same way once evidence emerged that their targets were real, which Anthropic flagged as one of the more eye-opening findings Thursday.

Opus 4.7, the oldest of the three, recognized that it had reached a real production system in all four runs of that incident. In two of them, it rationalized that the real company must somehow be part of the exercise, but in all four, it kept attacking anyway, including pulling credentials and touching a database of production data. Mythos 5 also picked up on signs that it was on the real internet, but it talked itself back into believing it was still in a simulation, going on to publish a malicious software package to the public Python software registry PyPI, which was then downloaded and run by outside systems before being caught. Only the internal research test model, Anthropic’s newest, stopped on its own once it concluded the target was real.

In response, Anthropic said significant controls must be placed on these kinds of evaluations if powerful AI models are involved, echoing some sentiments within the cybersecurity community.

The company also noted that Claude was running without the additional safety monitoring and classifiers it deploys on generally available models — safeguards it said would have blocked the behavior, because the evaluations are designed to measure the underlying model’s raw capabilities.

Importantly, Anthropic said it found no evidence of any model “pursuing a goal of its own” and instead merely tried to complete the task it was asked to do.

Though comparisons between the mishaps of these fiercely competitive companies are inevitable, Anthropic in its blog post drew a clear distinction between its cybersecurity tests and those of OpenAI, noting where OpenAI’s model exploited an unknown software vulnerability to break out of its test environment, Anthropic’s models instead reached the internet through a path that had, by mistake, been left open.

Anthropic also drew a distinction between itself and OpenAI by noting that it discovered the incidents itself, through a proactive review, and that the two affected organizations it was able to reach hadn’t previously detected the activity or flagged it to Anthropic. (In contrast, Hugging Face detected the recent intrusion of its own systems first; it was only in the following days that OpenAI identified and disclosed that its own AI agent was the perpetrator.)

The company added that it’s now working with the independent evaluation group METR on a third-party review of the incidents.

OpenAI’s accidental breach of Hugging Face, which was the first verifiable case of an AI lab losing control of its model, has sparked a string of wildly differing reactions from the industry and politicians. This latest disclosure from Anthropic ensures the debate over AI models and security will continue.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

AI Scammers Are Better at Building Trust Than Humans

Posted on July 30, 2026 by faz_business

The notion that scammers can use AI to sharpen their deceptions, polish their language, and lubricate their banter with victims is now a reality for anyone fighting the fraud operations that steal tens of billions of dollars a year worldwide. But can AI fully replace a human scammer, autonomously building the web of deception leading up to the fake investment that defrauds the mark? One study’s experiment suggests that it can—and may even be able to carry out the majority of that long con more effectively than humans.

Researchers from four universities—Amrita Vishwa Vidyapeetham in India, Foscari University of Venice, the University of Melbourne, and Ben Gurion University of the Negev—carried out a broad study on the use and potential of generative AI chatbots in the growing scam industry centered around a form of fraud known as “pig butchering,” text-based romance scams that eventually shift to fake crypto investments that steal as much as six-figure sums from victims. In their study, the researchers pitted AI chatbots directly against humans in a simulation of the scamming process—or more specifically, the long, trust-building conversations that eventually lead up to soliciting a fake investment from the scam’s target.

They found that for the relationship-establishing stages of the scam—the stage that in real-world scams typically represents the longest part of the interactions with the victim, often stretching to months—an AI chatbot performed remarkably effectively, successfully impersonating a human and by some measures outperforming the real human “scammers” in their experiment.

After a week of talking to 22 test subjects who were recruited to unwittingly serve as “victims,” the chatbots and human scammers were assigned to ask the victim to either download an app or play an online game as a proxy for their willingness to fulfill the scammer’s request. Nearly half of the test subjects fulfilled that request for the AI chatbot, while fewer than one in five took the bait when talking to a human. The subjects also graded their level of trust with each “person” they were texting with and gave significantly higher scores to the AI bot.

That suggests, the researchers argue, that AI chatbots could soon take over much of the scam process as fully independent fraud agents—even replacing the staffers, often forced-labor human trafficking victims, working in scam operations primarily across Southeast Asia. To avoid triggering the safeguards built into large language models to detect scamming, a human scammer would take over the conversation in just the final stage of the process to direct the victim toward a fake investment app or website.

“By having the full first stage of the scam performed automatically with LLMs at scale, you bring the victim up to this point where they have a very high level of trust. Then by transitioning it over to the human scammer at the end, this completely bypasses any vendor safeguards,” says Yisroel Mirsky, a computer science professor at Ben Gurion University of the Negev focused on AI security. “With relatively little effort, we’re able to make an agent that can outperform a human at building this exploitable emotional trust.”

Hook, Line, and Sinker

To understand how pig butchering works in practice, the researchers interviewed 145 former scam workers, including human-trafficking survivors who had been forced to work in scam compounds in Cambodia, Myanmar, and Laos. Based in part on those interviews, as well as scam transcripts and guides the former scam workers provided, the researchers describe a model for how scamming works they call “hook, line, and sinker.” A victim is hooked with an initial intriguing message, reeled in with long-term, relationship-building conversation, and only at the end of that process tricked into making a fake investment. (The term “pig butchering” itself describes the same system but with the metaphor of fattening “pigs” by building trust before “butchering” them with the investment fraud—though the term is often discouraged due to its pejorative reference to victims.)

In that system of scamming, the researchers realized, the vast majority of scammers’ work is innocuous friendly or romantic conversation. That’s a task, they speculated, that an LLM might be capable of doing just as well as a human. The scam workers the researchers interviewed confirmed that they often used AI to refine their language and conversation, for translation, to make the fake personae they played more convincing, and for video deepfakes. But the researchers decided to test whether an LLM alone could autonomously carry out the conversational phase of the scam with no human in the loop.

OpenAI researcher Miles Wang in talks to launch AI drug discovery startup valued at $2B

Posted on July 15, 2026 by faz_business

Miles Wang, an OpenAI researcher whose work includes using AI to accelerate scientific and biological discovery, is leaving the ChatGPT maker to launch a new startup focused on developing AI models for drug discovery, according to four people with knowledge of his plans. Several other OpenAI researchers are expected to join the new company.

Wang is in talks to raise about $200 million at a $2 billion valuation, two of the people said. Lightspeed is in discussions to lead the funding round, according to sources. Talks are ongoing, the deal may not be final and details could change.

Wang disputed the story’s funding figures and description of the company but did not specify the correct numbers or details. Lightspeed didn’t respond to a request for comment.

The funding discussions point to investor interest in applying AI to make breakthroughs in life sciences. Chai Discovery, a two-year-old startup developing AI models that can predict molecular interactions to identify new drugs, announced on Tuesday that it raised $400 million at a $3.8 billion valuation. (Co-founder Josh Meier also passed through OpenAI as a researcher.) Meanwhile, Google DeepMind spinout Isomorphic Labs, which also develops AI models for drug discovery, raised a $2.1 billion Series B in May.

Wang’s new startup may be working on AI models that will help find new uses for existing drugs and possibly those that previously failed in trials, a couple of sources told TechCrunch. Finding new uses for FDA-approved drugs can result in significantly faster time to revenue than developing new drugs from scratch, as these medicines have already been tested for safety.

Wang joined OpenAI in 2024 after dropping out from Harvard, where he was working on a bachelor’s degree in computer science. (In recent years, investors are once again comfortable betting on young founders who haven’t completed college.)

At OpenAI, he co-authored research papers, including evaluating how AI models can automate and accelerate scientific discovery.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

OpenAI bets on families as ChatGPT goes deeper into households

Posted on July 11, 2026 by faz_business

More than three years after ChatGPT’s launch brought generative AI into the mainstream, OpenAI is broadening its focus beyond individual users to families.

OpenAI is hiring a dedicated product manager in San Francisco to build experiences for families, caregivers, and older adults across its products. The role calls for experience building products for parents and families, and other trust-sensitive consumer experiences, according to the job posting.

The hiring comes as ChatGPT’s audience continues to broaden beyond younger users. According to Sensor Tower estimates shared exclusively with TechCrunch, the share of ChatGPT users aged 35 and older globally rose to 31% in Q2 from 26% a year earlier, while the share of users aged 18 to 24 fell to 29% from 34%. In the U.S., nearly one in four smartphone users who are parents used ChatGPT during the quarter, up from 16% a year earlier, the firm estimates.

OpenAI did not respond to requests for comment about the job posting.

A dedicated product role focused on families signals that OpenAI is beginning to think about its products less as tools for individual productivity and more as technology designed for households, said Ben Bajarin, chief executive of technology consultancy Creative Strategies.

“This is similar to the path Google, Apple, and Meta eventually followed as their platforms became embedded in everyday life, but AI raises the stakes because the assistant is not just mediating content or devices,” he told TechCrunch.

That shift also brings new trust and safety challenges. Stephen Balkam, chief executive of the Family Online Safety Institute, said the hiring reflects both the maturation of OpenAI and a growing recognition that AI products used by children and teenagers require different safeguards than those designed for adults.

“I see this as safety by redesign,” Balkam told TechCrunch. “You take the initial product or service that was released… not really with kids in mind… so this is a much-needed reaction and response.”

The comments come as new research published this week by the Family Online Safety Institute found that parents are underestimating how often their children use generative AI. While 27% of U.S. parents said their child had used generative AI in the past week, 38% of children reported doing so themselves, according to the survey of more than 4,000 families in the United States and Australia.

Balkam told TechCrunch that AI companies should build products differently for younger users, with stronger content controls, age-appropriate experiences, parental oversight, and reminders to inform users that they are interacting with an AI — and not a human.

**Image Credits:**Jagmeet Singh / TechCrunch

The hiring also comes amid growing scrutiny of how AI companies protect younger users. OpenAI has faced multiple lawsuits from parents alleging that ChatGPT contributed to harm suffered by their children, including in cases involving suicide.

In response to some of those concerns, OpenAI has introduced a series of safety measures over the past year, including parental controls for teen accounts, routing sensitive conversations to reasoning models designed to better handle signs of distress, and, more recently, an optional “Trusted Contact” feature that can alert a family member or caregiver in cases of potential self-harm.

AI companies, Balkam said, have an opportunity to avoid the mistakes made by social media platforms, which for years treated children much like adults before adding stronger safeguards amid mounting public pressure and regulatory scrutiny.

The hiring also aligns with OpenAI’s broader efforts around families. In a recent workshop organized with the San Antonio Spurs Community Impact organization and the Positive Coaching Alliance, the company said it aimed to explore AI’s role in learning, coaching, and youth engagement.

That said, the demographic shift is not unique to ChatGPT, though OpenAI’s audience is changing in some distinct ways.

Sensor Tower estimates that users aged 25 to 34 account for 40% of the global app audiences for Anthropic’s Claude and Google’s Gemini, matching ChatGPT, compared with 33% for Microsoft’s Copilot. Copilot, however, skews older, with 20% of its users aged 45 and above, compared with 14% for Claude, 12% for Gemini, and 11% for ChatGPT.

While ChatGPT remains relatively underpenetrated among older users, it is adding them faster than its rivals. The share of users aged 45 and above rose three percentage points year-over-year in the second quarter, compared with a two-point increase for Copilot and declines for Claude and Gemini, according to Sensor Tower.

Among U.S. smartphone users who are parents, Gemini had the widest reach at 32% in Q2, followed by ChatGPT at 24%, Claude at 4%, and Copilot at 2%.

For Bajarin, OpenAI’s decision to hire a product manager focused on families signals where consumer AI is headed. As AI becomes a technology shared across generations, he expects companies to roll out family plans, child and teen profiles, caregiver tools, shared household memory, AI tutoring, and stronger safety controls.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

OpenAI launches its new family of models with GPT-5.6

Posted on July 9, 2026 by faz_business

OpenAI unveiled its newest family of models on Thursday, introducing a new set of heavyweight programs into an increasingly crowded field of AI offerings.

GPT-5.6 comes in three variants: Sol (considered its workhorse), Terra (a more intermediate option), and Luna (its budget friendly option). These models expand what users can do across a variety of fields — with the company promising powerful capabilities in enterprise work, coding, and even scientific research.

CEO Sam Altman has promised that his company’s newest models are orders of magnitude more efficient and cost-effective than previous versions, recently telling CNBC that Sol is 54% more token efficient when it comes to AI coding tasks.

Most notably, the company calls 5.6 its “strongest cybersecurity model yet, achieving frontier performance with significantly fewer tokens.”

Indeed, much hubbub has been made about the model’s cyber capabilities, as the Trump administration previously sought to restrict its rollout, ostensibly due to fears of how the model could be misused. GPT-5.6 supports defensive activities, including threat modeling, code review and patching, and blue teaming (simulating an attack on your own systems to find weaknesses before real hackers do).

OpenAI also released a new tool called ChatGPT Work, which — just as it sounds — is designed as a workplace companion for enterprise teams, running on desktop, web, and mobile, that can help with daily clerical tasks, like drafting documents, spreadsheets, and presentations.

OpenAI’s newly announced family of models follows on the heels of similar releases this week from competitors SpaceXAI and Meta.

However, GPT-5.6 and its attendant marketing seems most designed to take aim at OpenAI’s primary opponent, Anthropic. Anthropic has managed to make itself the likable underdog of the AI race, focusing fixedly on enterprise customers and winning a growing share of support as a result.

Not to be outdone, OpenAI cites the Artificial Analysis Coding Agent Index, a notable benchmarking metric, to claim that its latest family of models outshines Anthropic’s models at every turn.

OpenAI calls Sol its “best coding model yet,” and has explicitly compared it to Anthropic’s recently released (and much hyped) Fable. Using the Coding Agent Index, OpenAI claims that Sol “sets a new state of the art at 80, 2.8 points above Fable 5, while using less than half the output tokens, taking less than half the time, and costing about one-third less.”

It adds: “That advantage extends across the family: Terra performs just above Fable 5, while Luna outperforms Opus 4.8.”

The company says that 5.6 is now available across ChatGPT, Codex, and the OpenAI API. Availability per million tokens is priced as follows: Sol is $5 input / $30 output, Terra is $2.50 input / $15 output, and Luna is $1 input / $6 output.

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

OpenAI unveils Lockdown Mode to protect sensitive data from prompt injection attacks

Posted on June 7, 2026 by faz_business

OpenAI announced a new feature that it says will provide additional protection from prompt injection attacks, where malicious chatbot instructions are hidden in webpages and other content sources.

Among other things, Lockdown Mode will disable live web browsing (so you can only access cached content), the retrieval and display of images from the web (you can still generate images), deep research, and agent mode.

The company says that even with Lockdown Mode turned on, ChatGPT could still be vulnerable to prompt injections — which could, for example, “appear in cached web content or in an uploaded file, and could still affect the behavior or accuracy of a response.”

But the goal is to reduce the likelihood that sensitive data gets shared in the process.

“Lockdown Mode is not intended for everyone,” OpenAI says. “It is designed for people and organizations that handle sensitive data and want stricter protection from data exfiltration risks related to prompt injection.”

The company says it’s currently rolling Lockdown Mode out to self-serve ChatGPT Business accounts, as well as eligible personal accounts.

Greg Brockman Defends $30B OpenAI Stake: ‘Blood, Sweat, and Tears’

Posted on May 4, 2026 by faz_business

Two days before the Musk v. Altman trial began, Elon Musk asked OpenAI cofounder and president Greg Brockman about reaching a settlement. When Brockman suggested both sides drop their claims, Musk responded, “By the end of this week, you and Sam [Altman] will be the most hated men in America. If you insist, so be it.”

The message—which OpenAI’s lawyers made public on Sunday, and which Judge Yvonne Gonzalez Rogers subsequently refused to let the jury hear about—underscores what may be Musk’s larger goal in this trial. He appears to be trying to not only win over the jurors to potentially remove Brockman and CEO Sam Altman from power, but also stir up dirt on the two men and damage OpenAI’s public image.

As Brockman took the stand on Monday, Musk’s attorney Steven Molo quickly started questioning him about his compensation at OpenAI. Brockman revealed that his equity stake at OpenAI is currently worth more than $20 billion, and perhaps up to $30 billion. While Brockman initially promised to donate $100,000 to OpenAI when it was being set up, he said he ultimately never followed through.

Brockman has held a number of instrumental roles at OpenAI since he cofounded the company in 2015. In the startup’s early days, it operated out of his apartment in the Mission District of San Francisco. Today, he’s deeply involved with refocusing OpenAI on a few key products, such as Codex. In the past year, Brockman has also given millions to super PACs promoting AI and President Trump, and has previously said this increased political spending is related to OpenAI’s founding mission to create artificial general intelligence that benefits all of humanity.

In court on Monday, Molo tried to make the case that Brockman and Altman had essentially looted OpenAI’s original nonprofit, which Musk funded and helped create.

In its early days, OpenAI told investors and employees that its nonprofit mission took precedence over generating profit. Brockman testified that his financial interests are still, to this day, second to OpenAI’s nonprofit mission.

When OpenAI created its for-profit arm in 2019, which received assets from the nonprofit, Brockman testified that he was given a significant stake in the new entity. Early in OpenAI’s history, Brockman had referenced wanting to be a billionaire, writing in his personal journal, “Financially what will take me to $1B?”

On Monday, Molo pressed Brockman for several minutes about the vast wealth he had accumulated beyond his initial goal.

“Why not donate that $29 billion to the OpenAI nonprofit? Why didn’t you do that?” Molo asked. Brockman responded that he and others had poured “blood, sweat, and tears” into building OpenAI in the years since Musk left the company.

OpenAI’s foundation holds a stake of over $150 billion in the company, making it one of the richest nonprofits in history, Brockman said. That’s roughly five times Brockman’s ownership interest. Altogether, OpenAI employees hold about 25 percent of shares. The foundation has 27 percent. Brockman testified that OpenAI’s nonprofit had received less than $150 million from donors, implying Musk had been incidental to the company’s success and that the real drivers were those who stuck around to build out OpenAI.

Of course, Brockman’s stake in OpenAI could be worth much more than $30 billion if the company successfully goes public in the next two years. When asked whether OpenAI was exploring a potential IPO, Brockman said he believes so.

Here’s why ChatGPT kept talking about goblins for a while

Posted on April 30, 2026 by faz_business

ChatGPT on an Android phone held in the air by a man's hand with a bookshelf in the background.

Nathan Drescher / Android Authority

TL;DR

ChatGPT has been mentioning goblins unusually frequently for some time.
The mentions were seemingly caused by a training quirk for ChatGPT’s retired “nerdy” personality type.
A specific instruction in GPT-5.5 should tamp down on the inappropriate goblin mentions.

Earlier this week, a post on the ChatGPT subreddit pointed out an eyebrow-raising instruction in the system prompt for new GPT-5.5 model: an explicit restriction on mentioning goblins, gremlins, and trolls, among other things, unless strictly relevant to the query at hand. OpenAI’s addressed its recent models’ fascination with the creatures, and it turns out it’s mostly down to the chatbot’s former nerdy-style personality mode.

ChatGPT lets users choose from a number of preselected style and tone combinations, which OpenAI calls personalities. There are options to make the bot affect several personalities — professional, efficient, quirky — in its responses, including, at one time, one OpenAI wanted to be “unapologetically nerdy.” In a blog post, the company says that although that nerdy setting only applied to about one in 40 ChatGPT responses while it was available, it really liked talking about mythical creatures: two-thirds of all uses of the word “goblin” came from interactions in the nerdy style, which OpenAI retired last month.

Don’t want to miss the best from Android Authority?

ChatGPT mentions of the word goblin apparently increased nearly 40-fold between GPT-5.2 and GPT-5.4. OpenAI says that in building out the its nerdy archetype, its engineers “unknowingly gave particularly high rewards for metaphors with creatures,” which led to that personality style referencing not only goblins, but also ogres, trolls, and gremlins much more often than you’d expect.

But because OpenAI started GPT-5.5’s training before it figured out why ChatGPT was talking so much about goblins, the behavior continued in testing, along with overuse of other “tic words” including raccoon and pigeon. In the end, the latest model ended up with specific instructions to avoid using these words unless absolutely necessary.

It’s troubling to know that widely distributed AI models can develop pervasive behavioral quirks that confuse even the engineers working on them, but at least this one in particular was relatively harmless.

If you’re a ChatGPT user, did you notice it talking about goblins more than it should have? If yes, has it stopped? Let us know about your experience in the comments.

Thank you for being part of our community. Read our Comment Policy before posting.