AI Health Advice? Doctors Warn of Dangerous Guesswork

Patients are increasingly asking artificial intelligence chatbots for medical advice, but doctors warn the technology can produce dangerous errors. AI tools like ChatGPT can answer health questions in seconds, with users describing symptoms and receiving instant suggestions. However, these systems often invent facts — a problem experts call "hallucination." The tools can also miss serious conditions or recommend harmful treatments [267673].

Health officials stress that AI chatbots are not doctors. They cannot examine patients, review medical history, or order tests. Several cases have emerged of people delaying proper care after trusting chatbot advice [267673]. Regulators are racing to set safety rules as AI spreads into medicine. Some hospitals now use AI to assist staff — but under strict supervision. For now, experts urge the public to verify any AI health advice with a qualified professional [267673].

The warnings come as AI's role in healthcare expands rapidly. A PwC study estimates AI will have an impact of more than $800 billion on global healthcare in the next five years, with about 75% of that coming from efficiency savings and the remaining 25% from new business models. AI's presence in healthcare is expected to rise from 15% to 30% [263583].

Leticia Rodríguez Vadillo, head of retail, consumer and health at PwC Spain, warned at a health conference in Madrid that AI could create deep inequality in public healthcare systems if its adoption is not properly coordinated. "AI can blow up equality in access and treatment," she said [263583]. She cited several problems facing health systems: unclear regulations on bias and ethics, low trust among doctors and patients, weak technology infrastructure, limited access to quality data, and a shortage of skilled professionals [263583].

Rodríguez Vadillo called for an end to isolated pilot programs. "We cannot allow the game of pilots: one here and another there, where one hospital does something but the one next door doesn't know about it," she said [263583]. She urged hospitals to include ethics, privacy and patient safety in their contracts from the start, and to create pre-contract testing spaces to ensure effective public purchasing [263583].

She stressed that technology requires changing care processes and training staff. "If we bring in AI but don't change processes, we are only adding a patch," she said [263583]. She noted that doctors often complain about spending 50% of their time on tasks they did not study medicine for, and said automating bureaucracy would return essential time to patient care [263583].

Rodríguez Vadillo concluded that the positive impact of digital transformation will not happen on its own. "AI will change things for us, but for it to change for the better, we must govern it and push it forward in an organized system," she said [263583].