AI Roundup: Your pharmacist is a robot, AI is more racist than your uncle, and an AI/EHR all-misery team

Artificial intelligence is rapidly advancing in health care and all facets of business and life. Here are the latest developments:

Domo arigato, Mr. Pharmaboto

Placing a call to a pharmacist might soon result in chirps, beeps, and a request for some oil. Pharmacy robots are being tested for order filling, freeing pharmacists up to explain to patients why it takes two hours to fill a prescription for a week’s worth of Amoxicillin. AI is being used to help pharmacists monitor drug interactions, manage inventory based on predictive analytics – and possibly generate responses as to why it takes two hours to fill that Amoxicillin prescription.

No more house parties, says AI rental cop

If you were planning on having a house party in an Airbnb to celebrate seeing 1,000 patients in a week, AI is going to stop you. Airbnb is using AI to scan for red flags such as renting a house in the same city in which you live to stop this debauchery.It also looks at other Sherlock Holmes-esque clues like if the rental is for one night and occurring on a day when parties are common, such as Halloween or New Year’s. Is this really AI, or just common sense? Only AI can answer that question.

Strange infection? Doctors can ask AI what it might be

Doctors who need help diagnosing patients with strange ailments can access a program that can help them identify what their patient is suffering from. Now, when that patient who visited Tibet rolls into your office with a strange rash and a blue tongue, you have a place to turn for answers.

AI teaming up with EHRs to make physicians even more miserable

Apparently, EHRs cannot be fixed by humans, so artificial intelligence has been assigned the task. Maybe it won’t be able to solve the big problems, but maybe it could start by not moving buttons and menus around every two months?

AI is more racist than your Uncle Joe

A study found that AI is full of debunked race-based biases and misinformation and could create risks for patients. Perhaps these large language models were programmed using family discussions from Thanksgiving when Uncle Joe goes on a racist rant after too many sips of brandy?