{"id":25074,"date":"2023-12-19T10:20:44","date_gmt":"2023-12-19T02:20:44","guid":{"rendered":"https:\/\/ljdevice.com.tw\/?p=25074\/"},"modified":"2023-12-19T10:20:44","modified_gmt":"2023-12-19T02:20:44","slug":"google-unveils-medlm-family-healthcare-focused-generative-ai-models","status":"publish","type":"post","link":"https:\/\/ljdevice.com.tw\/en\/google-unveils-medlm-family-healthcare-focused-generative-ai-models\/","title":{"rendered":"Google unveils MedLM, a family of healthcare-focused generative AI models"},"content":{"rendered":"

December 13, 2023<\/p>\n

\"\"<\/p>\n

Google thinks that there\u2019s an opportunity to offload more healthcare tasks to generative AI models \u2014 or at least, an opportunity to recruit those models to\u00a0aid<\/em>\u00a0healthcare workers in completing\u00a0their<\/em>\u00a0tasks.<\/p>\n

Today, the company announced MedLM, a family of models fine-tuned for the medical industries. Based on\u00a0Med-PaLM 2<\/a>, a Google-developed model that performs at an \u201cexpert level\u201d on dozens of medical exam questions, MedLM is available to Google Cloud customers in the U.S. (it\u2019s in preview in certain other markets) who\u2019ve been whitelisted through Vertex AI, Google\u2019s fully managed AI dev platform.<\/p>\n

There are two MedLM models available currently: a larger model designed for what Google describes as \u201ccomplex tasks\u201d and a smaller, fine-tunable model best for \u201cscaling across tasks.\u201d<\/p>\n

<\/div>\n

\u201cThrough piloting our tools with different organizations, we\u2019ve learned that the most effective model for a given task varies depending on the use case,\u201d reads a\u00a0blog post<\/a>\u00a0penned by Yossi Matias, VP of engineering and research at Google, provided to TechCrunch ahead of today\u2019s announcement. \u201cFor example, summarizing conversations might be best handled by one model, and searching through medications might be better handled by another.\u201d<\/p>\n

Google says that one early MedLM user, the for-profit facility operator HCA Healthcare, has been piloting the models with physicians to help draft patient notes at emergency department hospital sites. Another tester, BenchSci, has built MedLM into its \u201cevidence engine\u201d for identifying, classifying and ranking novel\u00a0biomarkers<\/a>.<\/p>\n

\u201cWe\u2019re working in close collaboration with practitioners, researchers, health and life science organizations and the individuals at the forefront of healthcare every day,\u201d writes Matias.<\/p>\n

Google \u2014 along with chief rivals Microsoft and Amazon \u2014 are racing desperately to corner a healthcare AI market that could be worth\u00a0tens of billions of dollars<\/a>\u00a0by 2032. Recently, Amazon launched\u00a0AWS HealthScribe<\/a>, which uses generative AI to transcribe, summarize and analyze notes from patient-doctor conversations. Microsoft is piloting various AI-powered healthcare products,\u00a0including<\/a>\u00a0medical \u201cassistant\u201d apps underpinned by large language models.<\/p>\n

\n

But there\u2019s reason to be wary of such tech. AI in healthcare, historically, has been met with mixed success.<\/p>\n

<\/div>\n

Babylon Health<\/a>, an AI startup backed by the U.K.\u2019s National Health Service, has found itself under repeated scrutiny for making claims that its disease-diagnosing tech can perform better than doctors. And IBM was forced to sell its AI-focused Watson Health\u00a0division<\/a>\u00a0at a loss after technical problems led customer partnerships to deteriorate.<\/p>\n

One might argue that generative models like those in Google\u2019s MedLM family are much more sophisticated than what came before them. But research has shown that generative models aren\u2019t particularly accurate when it comes to answering healthcare-related questions, even fairly basic ones.<\/p>\n

One study co-authored by a group of ophthalmologists asked ChatGPT and Google\u2019s Bard chatbot questions about eye conditions and diseases, and found that the majority of responses from all three tools were\u00a0wildly off the mark<\/a>. ChatGPT\u00a0generates<\/a>\u00a0cancer treatment plans full of potentially deadly errors. And models including ChatGPT and Bard\u00a0spew<\/a>\u00a0racist, debunked medical ideas in response to queries about kidney function, lung capacity and skin.<\/p>\n

In October, the World Health Organization (WHO) warned of the risks from using generative AI in healthcare, noting the potential for models to generate harmful wrong answers, propagate disinformation about health issues and reveal health data or other sensitive info. (Because models occasionally memorize training data and return portions of this data given the right prompt, it\u2019s not out of the question that models trained on medical records could inadvertently\u00a0leak<\/a>\u00a0those records.)<\/p>\n

<\/div>\n

\u201cWhile WHO is enthusiastic about the appropriate use of technologies, including [generative AI], to support healthcare professionals, patients, researchers and scientists, there\u2019s concern that caution that would normally be exercised for any new technology is not being exercised consistently with [generative AI],\u201d the WHO said in a statement. \u201cPrecipitous adoption of untested systems could lead to errors by healthcare workers, cause harm to patients, erode trust in AI and thereby undermine or delay the potential long-term benefits and uses of such technologies around the world.\u201d<\/p>\n

Google has repeatedly claimed that it\u2019s being exceptionally cautious in releasing generative AI healthcare tools \u2014 and it\u2019s not changing its tune today.<\/p>\n

\u201c[W]e\u2019re focused on enabling professionals with a safe and responsible use of this technology,\u201d Matias continued. \u201cAnd we\u2019re committed to not only helping others advance healthcare, but also making sure that these benefits are available to everyone.\u201d<\/p>\n

Source:Google unveils MedLM, a family of healthcare-focused generative AI models | TechCrunch<\/a><\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

December 13, 2023<\/p>\n","protected":false},"author":7,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[609],"tags":[],"class_list":["post-25074","post","type-post","status-publish","format-standard","hentry","category-industrial-news"],"_links":{"self":[{"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/posts\/25074","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/users\/7"}],"replies":[{"embeddable":true,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/comments?post=25074"}],"version-history":[{"count":2,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/posts\/25074\/revisions"}],"predecessor-version":[{"id":25077,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/posts\/25074\/revisions\/25077"}],"wp:attachment":[{"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/media?parent=25074"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/categories?post=25074"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ljdevice.com.tw\/en\/wp-json\/wp\/v2\/tags?post=25074"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}