The 2025 Physicians AI Report by Offcall represents one of the most comprehensive studies of clinician sentiment toward artificial intelligence to date. The study surveyed more than 1,000 practicing physicians within the United States. A critical strength of this dataset is its breadth; the respondents represent 106 distinct medical specialties. This diverse sample ensures the insights reflect the reality of modern medicine across various practice settings, from high-volume emergency departments to specialized outpatient clinics.

The "Silent Revolution" of Physician Adoption

The prevailing narrative in the healthcare industry suggests that physicians are the primary bottleneck to digital transformation. This report thoroughly debunks that myth. Far from being "Luddites," physicians are adopting AI at a pace that outstrips their own institutions.

The frequency of use is staggering: 51% of physicians use AI multiple times every day, while another 16% use it once or twice daily. In total, 89% of the workforce engages with AI at least once a week. Only 3% of the surveyed population reports never using AI at all. This adoption is not being forced by hospital mandates; it is happening organically because doctors find the tools valuable. Approximately 84% of respondents say AI makes them better at their jobs, with 36% stating it makes them "much better."

How Physicians Use AI: From Machine Learning to Generative AI

A major takeaway from the report is the transition from "Hidden AI" to "Active AI." In the past, physicians interacted primarily with Machine Learning (ML) models—predictive algorithms embedded in Electronic Health Records (EHRs) that flag sepsis risks or predict readmissions. These models are often perceived as "black boxes" that clinicians find intrusive or unhelpful.

In contrast, the current surge is dominated by Generative AI (GenAI), which physicians use as a "cognitive co-pilot." The reseach have mentioned a number of AI products used by physicians and every one of those products in GenAI-based.

• Clinical Knowledge Retrieval: The top-rated tool is OpenEvidence, used by 44.9% of surveyed physicians. It leverages GenAI to query vast medical databases and provide summarized, cited answers to clinical questions, acting as a more efficient version of traditional search engines.

• Administrative Automation: ChatGPT is the second most popular tool (15.6%), used primarily for "shadow work." This includes drafting insurance appeal letters, simplifying complex medical jargon for patient instructions, and summarizing long consultation notes.

• Documentation and Scribing: GenAI "ambient scribes," such as Abridge (4.9%), represent a high-growth category. These tools listen to the patient encounter and automatically generate a structured medical note. 65% of physicians identify the elimination of manual charting as their highest priority for AI, seeing it as the primary solution to "pajama time"—the hours spent documenting after the clinical day ends. 

The Great Misalignment: Organizational Failure

Despite the personal success physicians have found with AI, they are deeply frustrated with their employers. A total of 81% of physicians are dissatisfied with their organization’s AI adoption strategy. When asked to rate their employer's speed, 39% said it was "not good" and 42% said it was "just so-so." Only 19% of physicians are happy with the progress their organizations are making.

The frustration is rooted in a lack of agency. 71% of physicians feel they have "little to no influence" on the selection of AI tools that are eventually imposed upon them. This is a critical error for leadership: 67% of physicians say that having a voice in AI selection would directly improve their job satisfaction. Furthermore, communication is failing; 48% of doctors rate their employer’s transparency regarding AI as poor.

A Conflict of Interests: Hopes and Fears

Physicians are remarkably optimistic about the long-term impact of AI, with 78% believing that patients will be healthier in ten years because of these technologies. However, this optimism is tempered by fear—not of the AI itself, but of those who control it.

The report identifies a significant trust gap. Physicians are terrified that AI will be used by administrators and insurance companies to prioritize cost-cutting and "throughput" over patient care. There is a fear that if AI makes a physician 20% more efficient, the hospital will simply demand they see 20% more patients, thereby negating any relief from burnout.

Additionally, liability remains a major hurdle. 87% of physicians believe they should not be held legally responsible for errors generated by an AI model. Without clear malpractice protections and clinical agency, physicians worry that AI will become the "EHR 2.0"—a technology that was promised to help but ended up increasing the administrative burden.

Conclusion and Recommendations

The 2025 report serves as a wake-up call for healthcare leaders. Physicians are already using AI; they are paying for it out of their own pockets and integrating it into their daily workflows. 89% of physicians believe they should receive direct funding or stipends for the AI tools they need.

To win the trust of the medical workforce, organizations must pivot from "imposing" AI to "empowering" physicians to lead its deployment. The goal of AI should not just be operational efficiency, but the restoration of the patient-physician relationship by removing the digital barriers that currently define modern medicine. Success will be found when AI is used to give time back to the doctor, rather than just extracting more labor from them.