When will the true application of artificial intelligence appear? This is a million-dollar question. What people really want to know is why, if artificial intelligence is already so intelligent, it has not yet provided any specific solutions in the field of medicine. While OpenAI is still focused on ChatGPT, Google is trying to develop a medical chatbot, which is exactly what people expect.
According to a recent report from The Wall Street Journal, Google is testing an advanced artificial intelligence program specifically designed to answer medical questions. This AI tool is called Med-PaLM 2 and is a medical chatbot that provides secure answers to medical questions. Google is testing this product at renowned medical institutions such as the Mayo Clinic.
Med-PaLM 2 aims to use Google's LLM (Large Language Model) to accurately and securely answer medical questions by aligning with the medical field. It is based on Google's language model, PaLM 2.
According to a blog post released by Google in April, Med-PaLM 2 is the first LLM to achieve "expert" level performance on the MedQA dataset of USMLE-style questions, with an accuracy rate of over 85%. It is also the first artificial intelligence system to achieve a passing score on the MedMCQA dataset, which includes questions from the Indian AIIMS and NEET medical exams, with a score of 72.3%.
It is important to note that the application of artificial intelligence in the medical field is very complex and involves numerous ethical, legal, and regulatory factors. Ensuring patient safety, privacy, and the reliability of artificial intelligence in providing accurate medical advice are important aspects that need to be addressed.
Achieving the true application of artificial intelligence in the field of medical science requires extensive research, validation, clinical trials, and collaboration with healthcare professionals and regulatory agencies. It is a gradual process, and while progress has been made, the comprehensive implementation of specific solutions in the field of medical science may take time.
Is passing the medical exam enough?
Passing an exam does not necessarily mean you are an excellent doctor. The exam is just a standard for society to assess whether you have basic domain knowledge and whether patients can trust you. According to official data, 91% of candidates passed the first stage of the USMLE exam in 2022. Does this mean they are all excellent doctors?
A doctor's expertise comes from real-life situations, which vary depending on different patients. Patients have unique characteristics, and prescribing medication cannot be generalized. Each patient has different bodily functions, and individual factors such as allergies must be taken into account. Google recognizes this complexity and acknowledges the importance of personalized healthcare.
A research paper titled "Towards Expert-Level Medical Question Answering with Large Language Models" published by Google and DeepMind acknowledges the limitations of Med-PaLM 2. The paper states, "We note that our results should not be considered applicable to every environment and audience for medical question answering."
Med-PaLM 2 was trained using the MultiMedQA dataset, which includes multiple-choice and long medical question answering data, but does not include personal patient data, following ethical standards.
However, obtaining personal patient data would greatly enhance its efficiency, but it is likely that patients are unwilling to share their health information as it falls under personal privacy. In addition, Google executives have confirmed that customers testing Med-PaLM 2 will retain control over their data, which will be stored encrypted and not accessible to the technology company, and the program itself will not absorb this data.
Should we ignore medical LLM?
Despite its limitations, the application of LLM in the medical field cannot be ignored. It just needs to be handled properly because it concerns life and death.
The field of medicine is making significant advancements every day, and it is difficult for doctors to maintain their top-level expertise.
In addition to Google, Microsoft is also making strides in the field of medical LLM. This tech giant, which is the largest investor in OpenAI, partnered with health software company Epic in April to develop tools using the algorithms behind ChatGPT to automatically draft messages to patients.
Junaid Bajwa, Chief Medical Scientist at Microsoft, stated in an exclusive interview with Analytics India Magazine, "Medical knowledge doubles every 73 days." He attributed this to the abundance of information coming from various parts of the world related to different diseases and medical conditions.
Source: https://analyticsindiamag.com/google-takes-ai-healthcare-in-its-med-palm/
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.