Exploring ChatGPT's Role in Healthcare

When faced with complex cases, Adam Rodman, a physician specializing in internal medicine and a Harvard Medical School professor, turns to ChatGPT for support. He's transparent with his patients about consulting the AI, making sure not to divulge private medical details.

In a unique instance, a patient actively engaged in the ChatGPT conversation, leading to a dynamic exchange between all parties involved. Rodman is among a growing number of medical professionals deliberating over which AI solutions, if any, to integrate into their practices. Many opt for versatile models like ChatGPT, recognizing limitations in specialization that might result in inaccurate outputs.

The pursuit for specialized solutions sometimes drives the adoption of innovative medical technology startups. According to Rock Health data, in early 2025, such AI-driven medical startups attracted 62% of venture funding in the digital health sphere, amassing $6.4 billion, a slight increase from the previous year's $6 billion.

The infancy of generative AI implicates that numerous tools lack extensive real-world testing, prompting many hospitals to conduct thorough in-house trials. These evaluations focus on ethical considerations and the practical utility of the tools. Despite the growing interest, a significant portion of healthcare professionals remain hesitant, with a recent survey indicating a leap from 26% to 48% in AI tool adoption over a year.

A commonality among interviewed doctors is the employment of AI to automate tedious tasks like note-taking, though opinions diverge on AI's safety and utility in other aspects. ChatGPT finds diverse uses beyond simple tasks like drafting emails. Companies like OpenAI and Microsoft are asserting their AI's diagnostic potential, with the latter claiming superior accuracy in solving case scenarios compared to human doctors. Despite this, some practitioners prefer using chatbots without exposing critical information.

Integrating AI into Medical Workplaces

At Boston Children's Hospital, Chief Innovation Officer John Brownstein launched a HIPAA-compliant version of ChatGPT to handle protected health details safely. Approximately 30% of staff use it for varied tasks such as crafting recommendation letters and data analyses. Meanwhile, Brownstein collaborates with AI startups, though the criteria for engaging these tools are becoming more selective. The evolving landscape of AI coding is reducing reliance on large teams, enabling hospitals to develop proprietary solutions internally.

David Zhang of Somos Community Care highlighted a fundamental weakness in ChatGPT when it incorrectly diagnosed a prostate biopsy. This emphasizes the advantage professionals possess in refining these AI systems. As the availability of AI-enabled medical devices grows, with over 1,200 entries in the FDA list, there is a proliferation of both promise and challenge for healthcare professionals.

Embracing and Critiquing AI Technologies

While some embrace innovations like AI-powered note-taking devices to alleviate 'clinician burnout,' others critique the fragmentation caused by multiple AI platforms, as noted by dentist Divian Patel. He co-founded Trust AI to consolidate AI services, garnering $6 million in seed funding.

Banner Health's Alan Weiss finds the influx of AI tools overwhelming and maintains rigorous vetting standards for new technologies. Many hospitals are implementing AI for administrative efficiency, although the anticipated benefits do not always materialize. Some studies have revealed AI-generated communication requires more correction than manual efforts, tempering initial enthusiasm.

Nonetheless, ambient listening devices are widely regarded as transformative, easing the documentation load on clinicians and enhancing doctor-patient interactions. Despite the potential of such technology, disclosure regulations require doctors to inform patients about the use of recording devices in certain jurisdictions.

The Future of AI in Medicine

The medical sector continues to negotiate between AI's potential and its ethical and practical boundaries. Some professionals, like cardiologist Francisco Lopez-Jimenez, acknowledge AI's profound impact on their field, noting a historical precedence of adopting innovative solutions.

As more clinicians navigate these technologies, ensuring consistent and responsible application remains crucial. Faster diagnostics facilitated by AI could enhance patient throughput, but this efficacy must be balanced against the potential for unintended consequences, as errors, albeit infrequent, have devastating potential.