A personal experience that could have had devastating consequences has highlighted both the potential and pitfalls of artificial intelligence in healthcare. Gleb Tsipursky credits his own AI health tool with potentially saving his life by accurately identifying deep vein thrombosis (DVT), a serious condition involving blood clots, which can be fatal if left untreated.
Initially attributing symptoms to a muscle spasm, Mr Tsipursky sought chiropractic treatment, but his condition worsened. Turning to his bespoke AI tool, trained on his medical records, medications, lab results, and visit notes, he was alerted to the possibility of DVT, with the AI flagging the need for an urgent ultrasound scan. This timely guidance led him directly to an emergency department, where four blood clots were subsequently found in his left leg.
According to NHS data, DVT is a serious condition that often goes undiagnosed, and can be fatal if part of the clot breaks off and travels to the lungs, causing a pulmonary embolism (PE). As many as 1 in 5 people who develop PE do not survive. Mr Tsipursky's family history of PE added to the gravity of his situation.
While acknowledging the crucial role of medical professionals in diagnosing and treating DVT, experts agree that AI tools like Gleb's can play a vital part in guiding patients towards seeking timely care. Research from Harvard Medical School found that large language models were more likely than physicians to provide accurate diagnoses in emergency department cases.
However, the rapid adoption of AI in healthcare raises concerns about patient safety and efficacy. NHS Digital reports that around 14% of UK adults use AI chatbots for medical guidance instead of consulting a GP. This trend has experts warning that chatbots cannot replace human assessment and care, highlighting the need for rigorous regulation, testing, and clinical supervision to ensure AI systems in healthcare are used responsibly.
The integration of AI into healthcare is seen as a valuable tool that can support clinicians and patients alike, but its role should be clear: to augment professional medical judgment and care, not replace it. As with any new technology, it's essential to strike the right balance between harnessing innovation and protecting patient safety.