It is aimed that artificial intelligence, which has recently developed seriously and amazes everyone, is used in different fields to provide great benefits to humanity. The statements made by Google today also show that we can see these technologies in medicine soon.
The US technology giant touched on artificial intelligence in this year’s leg of the annual The Check Up event, which focuses on health services. The company said it has established partnerships for the use of artificial intelligence on topics such as ultrasound readings, medical language models and cancer treatments.
Google has partnered to use artificial intelligence on important issues from breast cancer to tuberculosis
Google thinks artificial intelligence could be important in reading ultrasound devices. It is predicted that technology can provide great benefits, especially in regions where there are not enough experts. Although access to these devices is quite extensive, experts are needed to analyze and interpret images.
The company says that these processes can be facilitated by identifying data such as early detection of breast cancer and gestational age in expectant mothers with artificial intelligence models. To make this a reality, a partnership was formed with the Kenya-based nonprofit Jacaranda Health. This collaboration will explore artificial intelligence-based ultrasound treatments for mothers and babies and examine how such technologies can support pregnant women.
Google has also started working with Taiwan’s Chang Gung Memorial Hospital. With this partnership, the possibility of artificial intelligence being able to detect breast cancer via ultrasound as an alternative to mammograms, which is not available in some regions, will be explored. In addition, the company has partnered with AI-assisted screening to detect tuberculosis early, prevent its spread, and help provide treatment.
Google’s health-focused language model has improved significantly, giving answers just like a doctor.
Apart from these, the company also provided information with a medicine-oriented language model. It was stated that the new generation model, called Med-PaLM 2, was 85% successful in medical examination questions at the doctor level. This is 18% better than the previous version’s score. Greg Corrado and Yossi Matias, executives in Google’s health division, also noted that the model not only answers multiple-choice questions, but also correctly answers open-ended questions.
Of course, we cannot say that an artificial intelligence model like ChatGPT will replace doctors. Google also emphasizes in its statements that the technology is not ready for use in real-world conditions. In addition, there are still deficiencies in the model in evaluations regarding criteria such as bias, scientific facts, and medical consensus; Therefore, it is stated that more work is needed.