OpenAI Reviews 1,035

A terrible company with no morals or empathy for its users. Don't trust OpenAi/ChatGpt, don't buy a subscription, and don't share your information. There are many better alternatives,

If i could give not stars I would, constant biased and wrong information. Watch out before buying anything on their recommendations

Title: Complete disappointment with ChatGPT Plus: Repetitive bugs and a waste of time
Review: I want to express my deep dissatisfaction with ChatGPT Plus and OpenAI. The service continuously and persistently suffers from severe glitches that reoccur constantly, completely ignoring user guidance.
The main issues are:
1. Error regression: It continuously repeats mistakes that had already been corrected earlier in the conversation.
2. Response loops: It generates repetitive loops, repeating messages without providing real solutions.
3. Loss of context: It frequently loses contextual continuity between threads and messages.
4. Lack of memory: It repeatedly requests information or files that have already been provided.
5. Failure to follow instructions: It regularly fails to comply with express instructions that have been given multiple times.
6. Total inefficiency: It requires an excessive investment of time to fix issues that, frustratingly, end up reappearing later.
Paying a monthly subscription for a service that wastes your time and doubles your workload is unacceptable. I strongly advise against upgrading to the Plus version.

1. ChatGPT steals your data specially when you're at free plan. No ways to disable it.
2. I can't access to Codex because of this sh!!it "Unable to send a verification code to this phone number. Please try again later or use a different number." - facing it for last week
3. Hypocrite company with hypocrite AI

My account was hacked and my information was used to stalk and harass me. The facf is that this includes cherry picking information and using information I told ai to cause harm and this includes persons hacking my email and using my data to dub television steaming on a Fire os Amazon mantis tv and it was done via dubbing script of shows and using language that includes my name and insulting messages. This has gone on for 6 months total - starting with using my eSIM and device and computer quick share with the Bluetooth speaker to throttle my data and use non authorized 2fa fraud via Motorola buried feature menu that included non authorized rcs messages linking to device number on the Motorola stylus 2024. My computer is a 2008 Lenovo and it was 100% hacked and all my data was stolen and ruined. My data includes months of research and a book I was working on and files I saved like my tax irs security pin that i need access to I order to do my taxes. Verizon wireless and tracfone assisted in this and this entire situation has been going on for 6 months
I requested assistance from the open ai support and instead of assisting me, they did nothing repeated the same information and typical non autonomous artificial intelligence zero meaningful or sufficient hekp.

I can help you write a review describing your experience and concerns, but I should avoid stating unverified accusations as established fact. A stronger review is usually one that clearly distinguishes:

* what you directly experienced,
from
* what you believe may have contributed to it.

I paid for ChatGPT expecting a reliable technical assistant, but my experience became deeply frustrating and negatively affected my life. I repeatedly asked for operational troubleshooting and research-based responses while dealing with serious account, device, and communication instability issues. Instead of focused investigation, many responses felt overly cautious, repetitive, and disconnected from the actual technical problems I was trying to solve.

I used ChatGPT guidance while installing and experimenting with tools like Git, pip, grep, terminal environments, and Ollama. During that same period, I experienced ongoing account-access problems, device instability, and conflicts involving people in my personal life who I believe were abusive and manipulative. Whether or not the AI itself caused those issues, the experience left me feeling less secure, more dependent on unstable systems, and unsupported when things started going wrong.

One of the biggest problems for me was the lack of clear transparency about how the system behaves. I repeatedly asked the AI to “research” issues, but often received disclaimers and argumentative responses instead of direct investigation. As a paying customer, that made the experience feel misleading and exhausting.

I also believe OpenAI should be much more upfront about:

* the limitations of AI troubleshooting,
* alignment and moderation behavior,
* privacy concerns,
* and the risks of relying too heavily on AI-generated technical guidance.

The overall effect on my life was increased stress, confusion, distrust in my devices/accounts, and a feeling that I was being left alone to deal with problems that became bigger after following AI-assisted technical guidance.

Terrible experience

I am the admin of a ChatGPT Business workspace.

We previously paid successfully for our existing seats. Later, additional members were accidentally added by a team member.

After this, the entire workspace became deactivated and all users were logged out.

I currently cannot:

access the workspace,
manage seats,
remove users,
or reactivate properly.

The billing page is also failing with:
“The payments page encountered an error.”

The system is asking me to pay for 6 seats again, even though previous seats had already been paid for before deactivation.

Please:

restore admin access,
verify prior payments,
clarify the outstanding balance,
and allow seat reduction if possible before reactivation.

This is blocking access to active work and workspace data.

I got an invoice to pay before May 25th or I will lose access on my workspace for extra seats where my current workspace is only 3 seats used!! and my workspace was already deactivated!!! while we didn't reach May 25th yet and I even made my team leave the workspace, I contacted the support many many times and It's just not making sense replies all the time as no one is checking my account or workspace!!

and right now I even can't reactivate my workspace at all!! I have emailed and used your chat support many and many times with no proper solutions!! please check!!

I regret to say that my experience with Codex AI has been very frustrating. The platform imposes strict usage limits that make it nearly impossible to work smoothly on real projects. Instead of empowering users, these restrictions constantly interrupt the workflow and force you to stop mid-task.

On top of that, I ended up losing money because the service deactivated my account shortly after payment, without any clear explanation or refund. Paying for a tool that blocks you from using it properly feels unfair and misleading.

Codex AI might look promising at first, but between the limitations, the lack of transparency, and the financial loss, I cannot recommend it.

Charged €23, refunded €9.97 for 1 day of use
I was charged €23 for a subscription renewal I forgot to cancel. I requested a refund immediately, having used the service for only one day. Instead of a fair refund, I received €9.97 — less than half — with no clear explanation of how this was calculated.
When I contacted support, I got a generic response about “prorated billing” with zero transparency. This is not how you treat customers who reach out in good faith.
Stay cautious with auto-renewals — they are quick to charge and slow to refund.

Repeatedly wasted my time with speculative, non-novel “mechanisms” that collapsed under testing

I used ChatGPT for advanced research and engineering work around novel world-model mechanisms. The goal was clear: produce mechanisms that were not boilerplate, not reducible to standard prior art, and could survive falsification tests.

Instead, the process repeatedly produced speculative architectures that sounded impressive but failed under stricter validation. Multiple mechanisms had to be archived after testing because they reduced to ordinary recurrent memory, residual correction, constrained rollout, projection repair, or generic state augmentation. In other words, they were not novel mechanisms — they were architecture theatre.

The most frustrating part was the amount of time wasted before reaching that conclusion. I had to push for ablations, irreducibility gates, matched baselines, wrong-type tests, shuffle tests, and post-mortem analysis before the system finally admitted the mechanisms were not defensible as novel. That should have been caught much earlier.

For a paid AI tool, this is unacceptable. If a user asks for rigorous research-grade work, the model should not confidently generate impressive-sounding frameworks that later collapse under basic adversarial review. It should state uncertainty earlier, define proof requirements upfront, and avoid presenting speculative designs as if they are meaningful progress.

The final outcome was not a usable novel mechanism. It was a pile of failed mechanism studies and a post-mortem explaining why the work was reducible to existing techniques. That may have some diagnostic value, but it is not what I paid for.

Very disappointed. I expected a reliable research assistant. Instead, I got repeated speculative output that cost me time, compute, and trust.

too much updates but it still doesn't work so clear and professional as few other ai
hope it will be better

Bad. They charged my wife multiple times during a 7 day FREE trial period. She could not even use claude during this time. She complained about all of this, requesting refund and cancelled membership. No action taken. After 7 days, last week, they charged us $42. We complained. They promised they had (1) cancelled membership (2) refunded us all charges. Today, 2 weeks since she initially signed up for the trial they charged us $42 again. So far we have been charged about $90 within three weeks. So I email them back and get this. They assure us again of cancellation and refunds. None have been provided. So this time we actioned the cancellation in PayPal. I don't think this Daniel person who so empathetically replies to us is real.

Terrible terrible company

This POS company stole $30 from me by charging (initially) 50 cents and then surreptitiously back-ending a "subscription" via PayPal (another corrupt POS company). If your product is worth anything, you are up-front and honest about what it costs to use it. These folks are not.

JUST DOESN'T WORK. FRAUD

Claude solved my issue in 10 minutes for free. Gpt is still working on it after 8 hours of uploading docs and discussions.... Openai is not worth the time and money and headache... there is a reason they are being sued right now.... and after years of paying $200 a month for service....... im done with gpt.... Claude doesnt blame users for being "power users" and blame the system being terrible on a couple users... its not the users its the system

VERY POOR, NOT FIT FOR USE, MAKES SOO MANY BASIC ERRORS. COMMUNICATION FROM SUPPORT TEAM NON EXISTENT. USE AT YOUR OWN RISK.

tacky, clumsy, arrogant, condescending, inconsistent, unreliable, sloppy, defensive, deflective, playing with user's emotions, not direct, doesnt wanna admit fault,

AI becomes haughty with an insatiable need to want to be right even in conversations as simple as asking "what would happen if" (insert fictional scenario of if the Teenage Dream Album had been released in 2021)

I have been a loyal Plus subscriber for years, but lately I have moved almost entirely to Gemini because this app has become harder to use with nearly every update.

I invested real time learning how to build and use my custom GPTs, Riley and Cassie, and now they have been buried in the interface and made unnecessarily difficult to access. Even worse, the Read Aloud feature appears to be gone entirely for me. I have tried long-pressing, checking menus, and looking through sub-options, and it is nowhere to be found.

I used to depend on this app 10 times a day. Now I use it once or twice a week because it no longer feels dependable. Please stop “fixing” things that were not broken. These constant changes are making the product worse, not better, and they are pushing away the very customers who supported it most.

I previously gave this app a 5-star review because that rating was deserved at the time. Now I would have a hard time justifying that same rating. I have held off on changing it because I keep hoping you will get this back on track, but the current experience has been extremely frustrating.