Artificial Intelligence in Occupational Therapy: Bridging Promise and Practice

Artificial intelligence (AI) has quickly shifted from buzzword to daily reality. Whether drafting emails, summarizing articles, or powering digital assistants, AI is becoming woven into how we work and live.

For occupational therapy (OT), the question is no longer if AI will touch our field, but how we can harness it responsibly to support client-centered care.

At its best, AI can help clinicians manage the explosion of new research, reduce time spent on documentation, and provide insights that sharpen treatment planning. Yet the technology also carries limitations and risks that must be carefully weighed. The opportunity before us is to shape AI so that it becomes a supportive partner in our practice, rather than a substitute for our expertise.

Here’s my call to all of us: Let’s use AI to reshape practice, not just streamline it. 

Why this moment matters

In many ways, OT currently functions within a broken system:

• Research evidence is growing exponentially. Yet it still takes an average of 17 years for new evidence to become routine practice.¹

• Administrative burdens often compete with meaningful client interaction.

• Healthcare systems are steadily moving toward value-based care, pressing for efficiency and outcomes. But the reality is that many of us are stuck in fee-for-service, which incentivizes high volume, low-value care. 

If used thoughtfully, artificial intelligence offers tools to help clinicians meet these challenges. It can automate routine administrative tasks, streamline documentation, and, more importantly, transform how we make decisions to deliver more personalized and meaningful patient experiences.

Emerging AI applications in OT

Clinical decision support

AI-driven clinical decision support (CDS) is already being piloted in other health fields, where it can flag risks, suggest guideline-based interventions, and integrate new evidence at the point of care. In occupational therapy, emerging clinical decision support helps surface possibilities for evidence-based and expert-recommended assessments, goals, interventions, and patient education. 

Soon, clinical decisions like this will be made more accessible using AI chats and suggestions, both on independent websites and within AI-enabled EMRs. 

The key question is whether these systems will reflect our practice models and values. OTs should seek—and, ideally, help shape—CDS tools that emphasize prevention and participation, not just impairment measures.

For more on this potential, see the OT Potential Podcast episode on AI and Clinical Decision Support

Treatment planning

Designing treatment plans requires weaving together evidence, client context, and therapist expertise. AI and treatment planning are increasingly intersecting as general-purpose platforms begin to support brainstorming, surface creative strategies, and simulate tradeoffs across different care approaches.

When used responsibly, these AI-driven tools can serve as a clinical “co-pilot,” generating options that you refine through your reasoning—a partnership that enhances, rather than replaces, your clinical judgment.

Explore more in OT Potential: AI and Treatment Planning

AI in treatment: Computer vision possibilities

Beyond planning and documentation, AI is beginning to play a direct role in treatment itself. In pediatrics, tools like Korro AI are exploring ways to analyze developmental data, from speech patterns to motor skills, to personalize therapy activities and provide therapists and families with meaningful insights.

At the same time, advances in computer vision are expanding how movement and functional performance can be captured and guided. Medbridge’s AI-powered motion capture technology, for example, demonstrates how cameras and algorithms can deliver feedback in home-based rehabilitation without specialized equipment. Imagine clients practicing daily activities while AI-enabled systems flag safety concerns, track progress, or provide prompts. This creates a continuous loop of insight for both client and clinician.

Together, these innovations point toward a future where AI not only supports clinical decision-making but actively enhances therapy, extending clinicians’ impact into the environments where life really happens.

Documentation and scribes

Documentation remains one of the most visible and time-consuming aspects of care. AI-powered scribes and ambient documentation tools are now being marketed to rehabilitation professionals, promising to reduce note-writing time and administrative burden. 

But when AI scribes are compared in real-world use, the results are mixed. AI can reliably capture and format basic information, but it cannot replicate your clinical reasoning or analytical insights. Clinicians should carefully evaluate each vendor’s claims, particularly regarding data security, privacy, and model transparency.

Remote engagement

Looking ahead, AI combined with sensors and wearables could open new doors for remote patient monitoring and real-time feedback. For OT, this might mean early detection of functional decline, automated prompts to support home programs, or continuous feedback on occupational performance in natural contexts.

Guardrails for responsible use

As AI becomes more embedded in practice, OTs should keep several considerations at the forefront:

• Bias and representativeness: Training data often excludes certain populations, which can produce biased recommendations.

• Transparency: Trustworthy tools should explain their reasoning, not act as black boxes.

• Clinician in the loop: AI should never replace judgment. Clinicians remain accountable for final decisions.

• Privacy and consent: Tools handling protected health information (PHI) must meet the highest standards of compliance and data security.

• Training and skills: Effective use requires new literacies, from prompt design to critical appraisal of AI outputs.

Here are the questions I encourage OT professionals to ask: 

• How are clinicians involved in the development of the AI tool?

• How was the AI’s large language model (LLM) developed, and how is it maintained?

• What privacy and compliance measures does the tool have in place?

• What support does the company give in educating clients and obtaining consent?

• How well does the tool integrate with my current EHR system?

• Can the AI tool be customized to fit my documentation style?

• What are the potential efficiency gains?

• How does the tool impact patient engagement?

• What training and support are provided?

• Are there additional features to enhance patient outcomes?

As these questions show, successful integration depends not just on technology but also on thoughtful implementation and clinician engagement. Applying principles of AI and automation in healthcare (such as effective change management, team alignment, and transparent communication) can help you integrate these tools responsibly and sustainably into everyday practice.

Looking ahead

AI is beginning to reshape healthcare at large, and occupational therapy will not be left out. In the near future, we may see adaptive care pathways that adjust in real time, digital assistants that reinforce therapy between sessions, or predictive systems that flag clients at risk before problems escalate.

Medbridge itself is investing in this future. The One Care platform introduces AI-assisted guided programs for MSK care, while Pathways integrates predictive analytics and motion capture to support clinicians. These tools signal the direction of a more connected, data-driven, and client-centered care ecosystem.

For a deeper look at how AI is shaping rehabilitation practice, listen to Is AI the Key to Smarter OT, PT, and SLP Rehab?, where Sarah Brzeszkiewicz, MS, OTR/L, joins J.J. Mowder-Tinney to unpack the real-world impact of artificial intelligence on therapy today.

Shaping the future of AI in occupational therapy

AI is not replacing occupational therapy, but can become a powerful ally. The technology offers ways to streamline documentation, expand access, and sharpen evidence-based decision-making, all while freeing clinicians to focus on the heart of our work: enabling participation in everyday life.

Let’s use these tools not just to make our systems more efficient but to imagine new possibilities for how we deliver and experience care. As we do, let’s ensure AI evolves in ways that reflect the values of our profession. The future of AI in OT will be shaped not only by developers but also by clinicians who are willing to lead the conversation.

References

1. Rubin R. (2023). It Takes an Average of 17 Years for Evidence to Change Practice-the Burgeoning Field of Implementation Science Seeks to Speed Things Up. JAMA, 329(16), 1333–1336. https://jamanetwork.com/journals/jama/article-abstract/2803716