What “The Pitt” is right and wrong about AI in medicine in “8:00”

People keep telling me to watch The Pittand now that the show is tackling generative AI in healthcare, I have to. As Mashable’s technical editor, I answer a lot of questions about generative AI, and I also spend a lot of time talking to people who are extremely excited about artificial intelligenceand people extremely hostile to AI.

So what is the latest episode of The Pitt Season 2 be right about AI in medicine, and what’s wrong?

How AI takes into account The Pitt episode “8:00”.

Credit: Warrick Page/HBO Max

AI wasn’t the focus of season 2 episode two, “8:00.” We also get to see a nun with gonorrhea in her eye, an unhoused man with a colony of maggots on the inside of his moldy arm, and a clear view of an erection that, of course, lasted more than four hours.

However, one of the storylines focuses on newcomer Dr. Baran Al-Hashimi (Sepideh Moafi)’s attempt to modernize the emergency room and integrate AI into patient care. It features an AI application that automatically listens to patient visits and summarizes relevant details in their records.

When an enthusiastic student says, “Oh my God, do you know how much time this will save?” “, Dr. Al-Hashimi has an answer: thanks to the new AI application, emergency doctors will spend 80% less time compiling records. Later in the episode, the good doctor tells Dr. Robby (Noah Wyle) that it will also allow doctors to spend 20 percent more time with patients.

So far so good, but the application immediately makes an error. He documents the wrong medication in the patient’s record, replacing it with a similar-sounding medication.

This does not take away from Dr Al-Hashimi’s enthusiasm. “Generative AI is currently 98% accurate,” she says. “Always read carefully and correct minor errors. It’s excellent but not perfect.”

AI transcription East really good (and it still makes mistakes).

Dr. Al-Hashimi claims that generative AI is 98% accurate. But is it really true?

Verifying Dr. Al-Hashimi’s claims is actually tricky, because it’s not entirely clear what she means. If it only refers to the AI ​​transcription, it is closer to the truth. This is a task at which generative AI excels.

Last year, a group of researchers conducted a systematic review of 29 studies measuring AI transcription accuracy specifically in healthcare settings. (You can read the full review, published in the Medical informatics and decision making BMC newspaper.)

Some of these studies have found accuracy rates of 98% or higher, and others have saved doctors a lot of time. However, these studies involved controlled, quiet environments. In multi-speaker environments with lots of crosstalk and medical jargon, such as in a crowded emergency room, accuracy rates were much lower, sometimes as much as 50%.

Yet rapid advances in multilingual models are constantly improving AI transcriptions. So we could be generous and say that Dr. Al-Hashimi’s claim is close to the truth, for the latest LLM models, in some contexts.

Generative AI is certainly not 98% accurate.

Let’s take the latest version of ChatGPT as an example. When the GPT-5.2 model was released a few months ago, Documentation published by OpenAI on the model’s tendency to hallucinate and provide false information.

According to OpenAI, its GPT-5.2 thinking model has an average hallucination rate of 10.9 percent. It is Really high, especially since OpenAI wants ChatGPT to help you with medical questions. (The company recently launched ChatGPT Healtha “dedicated experience in ChatGPT designed for health and wellness.”)

Now, when GPT-5.2 Thinking has access to the Internet, its hallucination rate drops to 5.8%. But would you trust a doctor who is wrong 5.8 percent of the time – and only when he can use the Internet to verify his work? And would you like your AI health app to be connected to the Internet?

Generative AI could one day be 98% accurate. But we are not there yet.

Generative AI cannot replace doctors.

Credit: Warrick Page/HBO Max

Dr. Robby and other characters have a conversation about their “gut” instincts in this episode. This is something that generative AI can’t replicate, and it’s one of the reasons why many people don’t want to see AI tools entirely replace human workers, whether in the arts or medicine.

The episode also spends a lot of time showing examples of empathy. The best doctors not only have encyclopedic knowledge of their specialty. They don’t just have good instincts. They are true healers, who know that holding a patient’s hand at the right moment can be just as important as making the right diagnosis.

To be fair, most AI health tools I’ve seen aren’t trying to replace doctors. Instead, they want to give healthcare professionals more diagnostic tools and save them time. As Dr. Al-Hashimi says several times in the episode, she wants to give doctors more time at the bedside of their patients.

AI can help doctors save time.

Radiology, or medical imaging, is one of the most promising applications of generative AI in healthcare settings. When radiologists at Northwestern University implemented a custom generative AI tool to help analyze X-rays and CT scans, their productivity increased by 40% — and without compromising accuracy.

I think even many AI skeptics would agree that this is a positive outcome, for patients and doctors.

The Pitt seems to make Dr. Al-Hashimi a villain — or, at least, a foil for Dr. Robby — but good medicine and generative AI are not necessarily in conflict. Like any tool, it can be extremely useful – or extremely dangerous.

The Pitt Season 2 is now streaming on HBO Max, with new episodes airing every Thursday at 9 p.m. ET.

Disclosure: Ziff Davis, the parent company of Mashable, filed a lawsuit in April 2025 against OpenAI, alleging that it had violated Ziff Davis’ copyrights in the training and operation of its AI systems.