A 91±¬ΑΟ study compared how well artificial intelligence models and human clinicians handled complex or sensitive medical cases. The findings outlined where the technology showed promise and what limitations need to be addressed before AI is unleashed on patients.<\/p>\n\n\n\n
With growing provider shortages, especially in rural states like Maine, the study highlights how AI could help clinicians. Lessons from past integration of technology, which at times failed to enhance care delivery, offer valuable guidance for AI developers as they work to integrate the technology into clinical practice.<\/p>\n\n\n\n
The study authored by <\/a>C. Matt Graham, associate professor of information systems and security management at the Maine Business School, analyzed 7,000+ medical queries from the U.S. and Australia, comparing AI-generated responses with those of human providers. The verdict? AI shows potential as a clinical support tool, especially in data-heavy or procedural scenarios, but struggles with emotional nuance, cultural context and consistency over time. Discussions about how to make the most of AI\u2019s possibilities will take center stage at the Maine AI Conference<\/a> this Friday. We’re bringing together tech and business leaders, educators, policymakers and researchers to explore how Maine can lead the way in responsible AI.<\/p>\n\n\n\n Read the full story on 91±¬ΑΟ News<\/a>.<\/p>\n\n\n\n Contact: Erin Miller, erin.miller@maine.edu<\/p>\n","protected":false},"excerpt":{"rendered":" A 91±¬ΑΟ study compared how well artificial intelligence models and human clinicians handled complex or sensitive medical cases. The findings outlined where the technology showed promise and what limitations need to be addressed before AI is unleashed on patients. With growing provider shortages, especially in rural states like Maine, the study highlights how […]<\/p>\n","protected":false},"author":445,"featured_media":50887,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_seopress_robots_primary_cat":"none","_seopress_titles_title":"","_seopress_titles_desc":"","_seopress_robots_index":"","_kad_blocks_custom_css":"","_kad_blocks_head_custom_js":"","_kad_blocks_body_custom_js":"","_kad_blocks_footer_custom_js":"","_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":"","spc_primary_category":65},"categories":[609,94,65],"tags":[],"class_list":["post-50885","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-impact-newsletter","category-research-news"],"featured_image_urls_v2":{"full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes.jpg",1400,788,false],"thumbnail":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-150x150.jpg",150,150,true],"medium":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-300x169.jpg",300,169,true],"medium_large":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-768x432.jpg",768,432,true],"large":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-1024x576.jpg",1024,576,true],"1536x1536":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes.jpg",1400,788,false],"2048x2048":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes.jpg",1400,788,false],"archive_9_5":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-360x200.jpg",360,200,true],"portrait_3x4":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-105x59.jpg",105,59,true],"image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-600x338.jpg",600,338,true],"image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-600x280.jpg",600,280,true],"image_25_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-600x168.jpg",600,168,true],"3col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-317x178.jpg",317,178,true],"3col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-317x148.jpg",317,148,true],"3col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-317x178.jpg",317,178,true],"4col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-423x238.jpg",423,238,true],"4col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-423x197.jpg",423,197,true],"4col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-423x238.jpg",423,238,true],"6col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-634x356.jpg",634,356,true],"6col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-634x296.jpg",634,296,true],"6col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-634x357.jpg",634,357,true],"8col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-846x476.jpg",846,476,true],"8col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-846x395.jpg",846,395,true],"8col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-846x476.jpg",846,476,true],"9col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-951x535.jpg",951,535,true],"9col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-951x444.jpg",951,444,true],"9col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-951x535.jpg",951,535,true],"12col-image_16_9":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-1268x713.jpg",1268,713,true],"12col-image_15_7":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-1268x592.jpg",1268,592,true],"12col-image_full":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-1268x714.jpg",1268,714,true],"post-thumbnail":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-320x180.jpg",320,180,true],"gform-image-choice-sm":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-300x300.jpg",300,300,true],"gform-image-choice-md":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-400x400.jpg",400,400,true],"gform-image-choice-lg":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-600x600.jpg",600,600,true],"umaps-featured-image":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-500x375.jpg",500,375,true],"umaps-icon-size":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-32x32.jpg",32,32,true]},"post_excerpt_stackable_v2":" A 91±¬ΑΟ study compared how well artificial intelligence models and human clinicians handled complex or sensitive medical cases. The findings outlined where the technology showed promise and what limitations need to be addressed before AI is unleashed on patients. With growing provider shortages, especially in rural states like Maine, the study highlights how AI could help clinicians. Lessons from past integration of technology, which at times failed to enhance care delivery, offer valuable guidance for AI developers as they work to integrate the technology into clinical practice. The study authored by C. Matt Graham, associate professor of information…<\/p>\n","category_list_v2":"AI<\/a>, IMPACT Newsletter<\/a>, Research News<\/a>","author_info_v2":{"name":"emiller","url":"https:\/\/umaine.edu\/research\/author\/emiller\/"},"comments_num_v2":"0 comments","taxonomy_info":{"category":[{"value":609,"label":"AI"},{"value":94,"label":"IMPACT Newsletter"},{"value":65,"label":"Research News"}]},"featured_image_src_large":["https:\/\/umaine.edu\/research\/wp-content\/uploads\/sites\/281\/2025\/06\/brains-bytes-1024x576.jpg",1024,576,true],"author_info":{"display_name":"emiller","author_link":"https:\/\/umaine.edu\/research\/author\/emiller\/"},"comment_info":0,"category_info":[{"term_id":609,"name":"AI","slug":"ai","term_group":0,"term_taxonomy_id":609,"taxonomy":"category","description":"","parent":0,"count":5,"filter":"raw","cat_ID":609,"category_count":5,"category_description":"","cat_name":"AI","category_nicename":"ai","category_parent":0},{"term_id":94,"name":"IMPACT Newsletter","slug":"impact-newsletter","term_group":0,"term_taxonomy_id":94,"taxonomy":"category","description":"","parent":0,"count":112,"filter":"raw","cat_ID":94,"category_count":112,"category_description":"","cat_name":"IMPACT Newsletter","category_nicename":"impact-newsletter","category_parent":0},{"term_id":65,"name":"Research News","slug":"research-news","term_group":0,"term_taxonomy_id":65,"taxonomy":"category","description":"","parent":0,"count":226,"filter":"raw","cat_ID":65,"category_count":226,"category_description":"","cat_name":"Research News","category_nicename":"research-news","category_parent":0}],"tag_info":false,"_links":{"self":[{"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/posts\/50885","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/users\/445"}],"replies":[{"embeddable":true,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/comments?post=50885"}],"version-history":[{"count":1,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/posts\/50885\/revisions"}],"predecessor-version":[{"id":50889,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/posts\/50885\/revisions\/50889"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/media\/50887"}],"wp:attachment":[{"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/media?parent=50885"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/categories?post=50885"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/umaine.edu\/research\/wp-json\/wp\/v2\/tags?post=50885"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\u201cTechnology is only one part of the solution,\u201d Graham says. \u201cWe need regulatory standards, human oversight and inclusive datasets. Right now, most AI tools are trained on limited populations. If we\u2019re not careful, we risk building systems that reflect and even magnify existing inequalities.\u201d<\/p>\n\n\n\n