The Google LLC logo can be seen at the Google Store Chelsea in New York,

Shannon Stapleton | Reuters

Google announced new healthcare initiatives and partnerships at its annual Check Out event on Tuesday. The Google Health team shared news about features coming to search, tools for building health apps, and the latest AI-powered health research.

At the event, the company discussed new partnerships for AI-assisted ultrasound, cancer treatment and TB screening, but the latest version of its medical large language model, called Med-PaLM, was the biggest buzz.

Google first introduced Med-PaLM late last year. It is designed to provide high-quality answers to medical questions. Med-PaLM was the first AI system to successfully achieve a passing score, or greater than 60%, on multiple-choice questions similar to those used in the US medical licensing exam.

The company said the second iteration of the technology, Med-PaLM 2, consistently performs at an “expert” level on medical exam questions. Med-PaLM 2 achieved more than 85% accuracy, 18% higher than previous results.

Dr. Alan Karthikesalingam, head of research at Google Health, said the company also tests Med-PaLM’s responses against those of real doctors and clinicians. Med-PaLM responses are evaluated for factual accuracy, bias and potential for harm, he said.

Karthikesalingam showed controlled examples – no live demonstration – of how Med-PaLM 2 can answer questions like “what are the early warning signs of pneumonia?” and “can incontinence be cured?” In some cases, the Med-PaLM 2 responses were the same or even more detailed than the clinician responses. But in other cases, Med-PaLM 2’s answers were not so accurate.

“You can see from this work that we are still learning,” Karthikesalingam said during the event.

Given the sensitive nature of medical information, Karthikesalingam said it may be some time before this technology is at the fingertips of the average consumer. He said it’s important to innovate responsibly and in a controlled environment.

Google will continue to work with researchers and experts on Med-PaLM, and Karthikesalingam said the company will share more updates on this in the future.

“The potential here is huge,” he said, “but it’s critical that real-world applications are explored responsibly and ethically.”

Source by [author_name]

Previous articleGPT-4 is bigger and better than ChatGPT, but OpenAI won’t say why
Next articleGPT-4 will make ChatGPT smarter, but won’t fix its flaws