Close Menu
Tech Nova Mindset – Empower Innovation and Forward Thinking

    Subscribe to Updates

    Get the latest creative news from FooBar about art, design and business.

    What's Hot

    AI Agents Are About to Flood the Workforce. No One’s Ready for It

    September 28, 2026

    When can we say AI made a scientific discovery?

    September 28, 2026

    How Delhi Cut Electricity Loss from 50 to 5 Percent

    September 28, 2026
    Facebook X (Twitter) Instagram
    Trending
    • AI Agents Are About to Flood the Workforce. No One’s Ready for It
    • When can we say AI made a scientific discovery?
    • How Delhi Cut Electricity Loss from 50 to 5 Percent
    • OpenAI Pauses Training Its Most Powerful Models After Rogue Agents Target Government
    • Who’s liable when AI agents go rogue?
    • Poetry for Engineers: User Interface Design
    • This Week’s Awesome Tech Stories From Around the Web (Through September 26)
    • Meta’s Muse Is Adults-Only. Why Does It Look Like a Kids’ Toy?
    Tech Nova Mindset – Empower Innovation and Forward Thinking
    • Home
    • Gadgets
    • Reviews
    • Tech News
    • Future Tech
    • AI & Robotics
    • How-To Guides
    • More
      • Cybersecurity
      • Startups & Innovation
    Tech Nova Mindset – Empower Innovation and Forward Thinking
    Home»Cybersecurity»OpenAI’s own support bot has no idea how ChatGPT works
    Cybersecurity

    OpenAI’s own support bot has no idea how ChatGPT works

    kirklandc008@gmail.comBy kirklandc008@gmail.comOctober 17, 2025No Comments6 Mins Read
    Facebook Twitter Pinterest LinkedIn Tumblr Email
    OpenAI's own support bot has no idea how ChatGPT works
    Share
    Facebook Twitter LinkedIn Pinterest Email

    ChatGPT’s support bot hallucinates a feature that doesn’t exist in the app.

    Screenshot by Tiernan Ray/ZDNET

    Follow ZDNET: Add us as a preferred source on Google.

    ZDNET’s key takeaways

    • The ChatGPT app’s support bot hallucinates the program’s features.
    • OpenAI currently has no in-app bug reporting, a serious omission.
    • Recent chats show the support bot has gotten more knowledgeable.

    One of the early successes companies have cited in implementing artificial intelligence in the enterprise is in customer-facing tasks such as customer support. 

    If OpenAI’s own automated customer support is the future, customers are in for a rough time. 

    Also: How to use ChatGPT: A beginner’s guide to the most popular AI chatbot

    OpenAI’s automated support function inside its ChatGPT app displays a surprising lack of knowledge of the features of the application, something I found out when I went looking for help with a bug in the app.

    (Disclosure: Ziff Davis, ZDNET’s parent company, filed an April 2025 lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)

    My quest began with a bug in the iPad Pro version of the ChatGPT app. I was using a subscription version of ChatGPT, the Plus version, for which I pay $21.78 a month. After several months of using the app without a problem, I suddenly noticed that resizing the window of the app caused it to freeze. 

    A small bug and a nightmare support chat

    On an 11-inch or 13-inch iPad Pro, running any version of iOS that can multitask, you can put the ChatGPT app into one of four possible window sizes, from a very narrow window to one slightly larger that is similar to the phone version of the app, to two larger sizes, including a full-screen window.

    I found that resizing the window to either full-screen or the size just below that, where a little bit of the Home Screen is visible, caused the app to become completely unresponsive. No touch functions generated any response, and even the interface elements of the app — text, borders, prompt field — were stretched as if the app could not refresh them to accommodate the change in size of the window.

    Screenshot by Tiernan Ray/ZDNET

    I was able to reliably recreate the problem each time I resized the app, and I was able to reliably duplicate the problem on a second iPad Pro.

    I went to the help section of the app, called Help Center, which you can find under your account information in the lower left corner of the sidebar of the app. None of the help articles were any help, so I decided to try the chat function that pops up in Help Center.

    Also: A minority of businesses have won big with AI. What are they doing right?

    The automated bot, powered by ChatGPT, suggested several steps that didn’t help, such as forcing the app to quit and reopening it, and deleting the app and installing it again from the Apple App Store.

    And, then, the support bot suddenly began to hallucinate something that doesn’t exist in the app. 

    “Send Feedback in the App,” was its suggestion. “You can report this bug directly from within the ChatGPT app (usually under Account or Support > Report a Problem/Send Feedback). This helps the developers get logs and may speed up a fix.”

    “Usually,” perhaps, but not in this case. There is no part of the Help Center that is “Report a problem” or “Send Feedback.” It doesn’t exist.

    There is a function called “Report,” which can be selected inside the current chat, from the menu in the upper-right corner, that is not for bug reports. That function is for reporting issues with the current chat, such as, “Violence and self-harm,” or “illegal activity.”

    Also: How people actually use ChatGPT vs Claude – and what the differences tell us

    I pointed this out to the bot, which, in the agreeable tone for which ChatGPT is famous, conceded it was not really a means of bug reporting. “Thank you for pointing that out — you’re correct. The ChatGPT iPad app currently only allows reporting individual messages or chats.” 

    (By the way, you can see the history of your support chats from the chat window in Help Center if you log in again from the Help Center. Best to log in before you engage in any support chats, so that you can recover a record of the support chat later.)

    Screenshot by Tiernan Ray/ZDNET

    Automated email support was just as bad

    I then emailed the support function from the OpenAI developer forum, support@openai.com, and described my problem again. 

    In a series of automated emails, after describing the issue, the support function repeated the same advice as the in-app support bot, including telling me to “report this bug directly from within the ChatGPT app (usually under Account or Support > Report a Problem / Send Feedback),” adding,  “This helps the developers get logs and may speed up a fix.”

    When I protested, the automated email again conceded the error, telling me, “Thank you for pointing that out — you’re correct.”

    Screenshot by Tiernan Ray/ZDNET

    I contacted OpenAI — a human spokesperson — and received confirmation via email that there is no bug reporting function:

    We don’t currently offer a dedicated consumer-facing bug reporting portal for the ChatGPT app. People can find troubleshooting steps and ways to share feedback through our official Help Center at help.openai.com.
    For security-related issues, we have a formal bug bounty program through Bugcrowd: https://bugcrowd.com/engagements/openai. You can also give feedback directly in the product — for example, by giving a thumbs down on responses that aren’t helpful or by reporting conversations that don’t meet your expectations. We’re working to make the automated support replies more accurate and effective. (There is a “bug bounty” for developers, but that’s not the same as bug reporting or feature requests by average users.)

    The lack of bug reporting is surprising, as it’s a standard offering with most software. 

    Also: Even the best AI agents are thwarted by this protocol – what can be done

    The more amusing fact is that OpenAI’s bot, presumably trained on the company’s vast corpus of development documents, didn’t have the correct facts about the app’s features.

    If this is the state of customer help via AI, it’s not only not a solution, but it makes the user’s frustration even worse.

    A small glimmer of hope?

    In my tests, the original bug persists with the latest app release, version 1.2025.280. Through experimentation, I found that not maximizing the app window is the only answer. 

    Also: I built a business plan with ChatGPT and it turned into a cautionary tale

    Interestingly, in my follow-up tests, the app’s support bot no longer hallucinates that I can report a bug. Instead, it states up-front that there is no way to report a bug in the app and asks for a description of the problem that can be forwarded to the developers. The bot has finally learned something about the app it’s living inside of.

    At least that’s progress, I suppose. 

    Get the morning’s top stories in your inbox each day with our Tech Today newsletter.

    bot ChatGPT idea OpenAIs support Works
    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
    kirklandc008@gmail.com
    • Website

    Related Posts

    Can AI reason without words? A small model puts the idea to the test

    September 22, 2026

    The Download: AI’s trillion-dollar gamble and OpenAI’s biology data bid

    September 16, 2026

    The Download: OpenAI’s turning point for math and a battery record

    September 9, 2026
    Leave A Reply Cancel Reply

    Top Posts

    Nothing CEO says phone prices are going to keep going up

    June 12, 20267 Views

    The best VPN routers of 2026: Expert tested and reviewed

    June 14, 20263 Views

    Google DeepMind Plans to Track AGI Progress With These 10 Traits of General Intelligence

    March 21, 20263 Views
    Stay In Touch
    • Facebook
    • YouTube
    • TikTok
    • WhatsApp
    • Twitter
    • Instagram
    Latest Reviews

    Subscribe to Updates

    Get the latest tech news from FooBar about tech, design and biz.

    Recent Posts
    • AI Agents Are About to Flood the Workforce. No One’s Ready for It
    • When can we say AI made a scientific discovery?
    • How Delhi Cut Electricity Loss from 50 to 5 Percent
    • OpenAI Pauses Training Its Most Powerful Models After Rogue Agents Target Government
    • Who’s liable when AI agents go rogue?

    AI Agents Are About to Flood the Workforce. No One’s Ready for It

    September 28, 2026

    When can we say AI made a scientific discovery?

    September 28, 2026

    How Delhi Cut Electricity Loss from 50 to 5 Percent

    September 28, 2026

    OpenAI Pauses Training Its Most Powerful Models After Rogue Agents Target Government

    September 28, 2026
    Facebook X (Twitter) Instagram Pinterest
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms and Conditions
    • Disclaimer
    © 2026 TechNovaMindset. Designed by By Pro.

    Type above and press Enter to search. Press Esc to cancel.