FAQ

Effective September 1, 2026, the AI Policy expanded to include education activities and a broader scope for research. The UW Medicine Innovation Core website includes a dedicated Policy Updates page that outlines the changes in the most recent version.

We are working towards an enterprise-wide AI policy that applies to all of UW Medicine inclusive of our training activities and those faculty, staff and learners who support our educational mission. Use of AI in education, including use in the clinical learning environment, is now included in the UW Medicine AI policy. We are committed to preparing our learners to become thoughtful, skilled and compassionate clinicians and researchers who can use emerging technologies responsibly and effectively in patient care, research and other professional settings. This policy also provides clarity regarding if and when AI can be used in the learning environment and what process and requirements apply to use in those settings.

As AI becomes more integrated into healthcare, it is important to establish clear standards that support both innovation and learning. AI should strengthen education — not bypass the development of independent human judgement, critical thinking, communication skills, professionalism and other core competencies that learners must develop. Our approach is intended to ensure all learners gain the critical thinking and analytical skills necessary to perform innovative research and provide high-quality patient care while also understanding how to effectively and responsibly use the newest technologies.

The UW Medicine AI Policy Glossary defines AI as “systems that apply AI algorithms and/or models to perform specific tasks that typically require human intelligence, such as visual perception, speech recognition, decision-making and language translation.” Additionally, AI systems can perform tasks beyond individual human capabilities, such as processing large datasets at high speeds. You can refer to Section A: General Requirements of the AI Policy for the AI uses that fall under the AI policy. If you have questions on a specific use case, please contact UWMInnovationCore@uw.edu.

UW Medicine data includes any data used by UW Medicine for any clinical, education, business operations or research purpose, regardless of origin. This extends well beyond protected health information (PHI) and other sensitive data types such as personally identifiable information (PII). Please see the AI Policy Glossary for full UW Medicine data definition.

Individuals using AI are responsible for identifying, reviewing and complying with applicable laws and policies. This includes verifying the output of any AI tool to assess its accuracy, appropriateness and potential biases. Individuals are accountable for any errors resulting from the use of AI tools.

The AI Workforce Literacy Task Force, which sits within UW Medicine’s AI governance structure and works in close collaboration with Enterprise Learning Office, is developing an enterprise-wide training and literacy strategy to ensure that our workforce understands how to use AI tools appropriately, safely and responsibly.

Yes, you may use publicly available AI tools if no UW Medicine data is exposed to the tool and the use is otherwise permitted by UW Medicine AI policy. Do not enter UW Medicine data (as defined in the AI Policy Glossary) into any AI tool unless UW Medicine has approved the tool for the specific data and intended use.

Yes, however, before purchasing or using an AI software or subscription, you must follow applicable procurement, contracting, privacy, security and records management processes and requirements. Additionally, depending on the intended use, data involved, or research activities, the AI software or subscription may require AI internal review.

Purple is UW’s in-house AI platform for writing, learning, analysis, coding and brainstorming. It provides access to approved AI models in a UW-managed environment.

UW Medicine employees are encouraged to use UW’s Purple as an alternative to other publicly available tools. However, at this time Purple can only be used for purposes that do not involve UW Medicine data.

UW Medicine is piloting and evaluating Purple for potential use with UW Medicine data. In the meantime, UW Medicine Chat remains the only AI chat tool approved for use with UW Medicine data.

Yes. If you are using an AI tool to record someone’s voice, image, or likeness, you should clearly inform the individual in advance and obtain any required consent before recording.

Washington State law (RCW 9.73.030) generally requires consent from all parties before recording private communications or conversations. This requirement applies regardless of whether AI is involved. Because AI tools may record, transcribe, analyze, or store communications, users should be transparent about how the tool is being used and obtain appropriate consent before recording.

Even when consent is obtained, users are still responsible for following UW Medicine’s AI Policy and considering privacy, security and bias risks associated with the AI tool.

For instructors, transparency is also recommended and considered best practice when using AI tools in instructional design, handling learner data, or evaluating learner work.

In addition, the Family Educational Rights and Privacy Act (FERPA) applies to higher education and graduate trainees. In general, students age 18 or older must consent before their education records are disclosed. Learner data may only be used with AI tools that have been approved by UW Medicine and used in accordance with applicable policies and acceptable use requirements. FERPA-protected data should not be entered into external AI tools unless explicitly permitted through approved institutional processes and policies.

Individual users of AI tools are responsible for validating AI outputs and are accountable for outcomes resulting from AI-generated errors. This includes use of AI for clinical care or within the clinical learning environment. Clinicians and learners must maintain all standards of practice and supervision articulated in UW Medicine Policies. If a medical error occurs due to the use of AI, the error should be reported and/or documented in the same manner as any other medical error.

AI tools offer significant opportunities but also pose risks and challenges, requiring careful evaluation of their benefits alongside potential drawbacks, particularly in areas like data privacy, bias, accuracy and reliability. UW Medicine has established AI governance to ensure responsible, ethical and equitable use of AI.

UW Medicine’s AI governance includes various representatives from key areas of expertise across UW Medicine to ensure broad participation while operating efficiently. Unfortunately, to ensure expediency in thorough reviews of use cases, not everyone can be included on these committees. However, we are actively developing an engagement strategy to involve individuals across UW Medicine in our innovation efforts. More information on opportunities to participate will be made available through our various communication pathways including Vitals, the bi-monthly Innovation Core Newsletter (“The Roundup”) and our new UW Medicine Innovation Core website.

UW Medicine works closely with the broader University of Washington community on AI initiatives, fostering knowledge-sharing and innovation while focusing on healthcare-specific needs. For AI use cases being considered for institution-wide implementation, UW Medicine partners closely with UW-IT to understand and explore UW Medicine’s participation. Additionally, research partnerships are being formed across various schools, institutes and centers. Given the unique regulatory demands of healthcare, medical education and research, UW Medicine has a comprehensive, centralized governance framework to ensure compliance with regulations, safeguard patient privacy and address risks unique to use of AI in a complex, academic healthcare setting.

Fred Hutch has its own committee structures and policies that govern the use of AI. UW Medicine and Fred Hutch are collaborating on several AI initiatives where appropriate and remain committed to exploring further opportunities for partnership.

Over the past few years, we’ve reduced the number of our data centers from seven to three. By the end of FY26, we will be down to two data centers. Within the next three years, we’ll be down to one plus the cloud. Cloud providers’ power usage effectiveness and carbon per unit of compute are significantly lower than hospitals can achieve on their own, and these companies are investing heavily in renewable energy, water-saving cooling technologies and carbon offsets. In addition to optimizing its physical infrastructure, UW Medicine employs a selective process when evaluating and adopting AI tools, choosing only those solutions that demonstrate clear value and efficiency. By optimizing infrastructure, adopting innovative practices and thoughtfully assessing each AI use case, UW Medicine is ensuring that the adoption of AI aligns with its commitment to reducing environmental impact while leveraging technology to enhance patient care and operations.

At UW Medicine, we are embracing the strategic use of artificial intelligence (AI) to advance our mission to improve the health of the public. From helping clinicians and researchers detect signs of disease in medical images to supporting patient care environments and streamlining time-consuming operational tasks, AI is enhancing the services we provide and the well-being of our employees. It will play an important role in UW Medicine’s future.

These technologies offer significant opportunities to improve care, expand access, accelerate discoveries, support the learning environment and streamline day-to-day work. In addition, we are committed to preparing our learners to become thoughtful, skilled and compassionate clinicians and researchers who can use emerging technologies responsibly and effectively in patient care, research and other professional settings. Realizing this potential requires both active engagement and careful stewardship. Together, we can shape this future and stay true to our values, by ensuring every tool is deployed in a responsible, ethical and equitable way across the entire organization.

Thesis/Dissertation/Meta Project/Capstones/General Exams & Proposals: If learners wish to use AI tools to assist with writing their degree milestones, they must discuss this with their chair, supervisor, or mentor and clarify expectations in advance of beginning their work. Academic programs may additionally set standards for using AI tools for writing these degree milestones.

Manuscripts: Journals have variable policies specifying limits and attribution standards for AI-assisted writing and analysis. Learners should consult journal policies before writing and submitting their manuscripts.

Grant Proposals: Sponsors/funders have variable policies regarding acceptance of AI-written grant proposals; if learners are planning to submit a proposal to the National Institutes of Health (NIH), they should consult NIH’s notice, Supporting Fairness and Originality in NIH Research Applications: NOT-OD-25-132.

Peer Review: Grant proposals and manuscripts shared as part of peer review should be treated as confidential documents. Most journals/sponsors have policies regarding their submission to AI tools, and learners should consult and ensure they follow those policies when engaging in peer review, as well as UW Medicine policies.

For clinical care, education or UW Medicine business operations:

• UW Medicine data: Only UW Medicine-approved AI tools, in accordance with their acceptable use documentation, can be used for handling UW Medicine data.
• Non-UW Medicine data that is not publicly available: Learners should consult the list of UW-licensed IT tools provided on the UW IT UWare website. These tools have enterprise data protection and are approved for use with data up to Level 3 per the UW Data Classification Scheme.
• Publicly available data sets, not considered to be UW Medicine data: Any AI tool can be used to handle these data.

Some research studies may have an exception under the AI Policy, if the AI tools are used only for data analysis, provided that:

• The research study is not designed to develop, validate or refine an AI system intended for use in clinical care, patient treatment, to support clinical decision-making, or other high-impact decision-making about individuals (e.g., eligibility for healthcare or insurance, employment decisions, criminal justice decisions, admissions, scholarship decisions).

• Access to models trained using UW Medicine data will not extend beyond the UW research team.

However, research studies will still be required to submit an IRB application and will require a UW Medicine security review and validation for any use of UW Medicine data that will be stored or retained outside a secure UW Medicine environment. For more information see Guidance for School of Medicine Research Involving AI.

Tools like OpenEvidence and UpToDate to retrieve and synthesize evidence can be used only if UW Medicine data is not put into the tool, and both the UW Medicine AI Policy and all other applicable laws and policies are followed. You must vet all AI-generated references and validate outputs for accuracy. Also, learners who use AI tools in completing assessments should disclose this.

For clinical care, education or UW Medicine business operations:

Review is required when the proposed use of AI has any of the following characteristics associated with possible elevated risk as outlined in Section B.1 of the AI Policy:

1. Its output is patient-facing without validation by an individual with the expertise necessary to detect an error (e.g., hallucinations, missing information).
2. A patient will be directly interacting with it.
3. It is trained on, has access to or individuals could input identifiable clinical data (including PHI), including limited data sets.
4. It is trained on, has access to or individuals could input UW Medicine De-identified health information that could be accessed by a third party (e.g., vendor or otherwise) or could leave the secure UW Medicine environment and has not been certified as de-identified through the honest broker process.
5. It is trained on, has access to or individuals could input UW Medicine data (e.g., sensitive data such as legal, financial, IP, PII, confidential/proprietary, trade secret, subject to contractual sharing limitations, academic or learner data protected under FERPA). (See AI Policy Glossary for UW Medicine data definition)
6. It affects and/or automates clinical care (e.g., diagnostics or decision-making) with or without human validation of output.
7. It affects and/or automates clinical coding/billing with or without human validation of output, and/or is designed to impact clinical coding/billing.
8. It fundamentally changes the work functions of represented (unionized) staff or trainees.
9. It requires significant financial investment, exceeding $100,000 (either one-time cost or annually, including internal labor).
10. It affects the selection of candidates for academic appointments, admission, scholarships, awards, or other forms of recognition.
11. It presents equity and/or accessibility impacts (e.g., language access, translation accuracy, digital-divide considerations and other impacts as determined when reviewing with the support of the Equity Impact Review Tool).
12. It involves a medium or high level of organizational change management to be successful.

For Research:

Review is required when research is conducted by a UW principal investigator (PI) that involves the following: (1) the use of AI, and (2) requires UW Institutional Review Board (IRB) review, and (3) either UW Medicine data or targeted enrollment of UW Medicine patients, staff, faculty, or learners.  See Section C of the AI Policy.

A determination that no internal review is required does not waive other applicable IT, privacy, security, procurement, contracting, research, education, operational approvals or an individual’s responsibilities when using AI tools.

To learn about how to submit a use case for review, visit the Innovation Core AI Intake page.

Internal review is not required for AI uses considered “low risk” under Section D of the AI Policy. Examples include UW Medicine approved AI tools used within their intended scope and limited uses of public AI tools that do not involve UW Medicine data or elevated-risk activities.

While low-risk use cases do not require internal review, they must adhere to other existing UW Medicine policies and processes (e.g., IS Risk Management Standard, Privacy, Confidentiality, & Information Security Agreement (PCISA)). Individuals are responsible for reviewing AI-generated outputs before relying on or acting on them to help identify and report potential risks, errors, bias or unexpected behavior.

For certain activities, learners must also receive authorization from their program, supervisor, or instructor prior to using AI tools, including UW Medicine Approved Tools or AI uses otherwise considered low risk. For more information on learner authorization requirements, see the Education Review – For Learners section in these FAQ below.

See Section D: “AI Uses Not Requiring Internal Review” for more examples and guidance.

For clinical care, education or UW Medicine business operations use cases:

• Submit the UW Medicine IT “New Project Idea” request form.

For research use cases:

• Complete the Human Subjects Research Determination Worksheet to determine if a research study involving AI requires internal review by the UW IRB. If the worksheet indicates UW IRB review is required, submit a UW IRB Application through Zipline along with the Artificial Intelligence Supplement Form.

Please note that not all uses of AI require internal review. See questions above or review the AI Policy to determine whether internal review is needed for your specific use case.

As part of the internal review process, each use case that is considered to have elevated risk is evaluated to assess the potential equity impacts of the tool. This evaluation helps identify potential risks related to bias (to the extent possible) that need to be addressed before moving forward. When selecting a new vendor for an AI purpose, UW Medicine ensures that vendors have a reporting pathway in case bias is found within the product.

Low-risk use cases (as defined under the UW Medicine AI Policy) are not internally reviewed but must adhere to other existing UW Medicine policies and processes (e.g., IS Risk Management Standard, Privacy, Confidentiality, & Information Security Agreement (PCISA)). Individuals are responsible for reviewing AI-generated outputs before relying on or acting on them to help identify and report potential risks, errors, bias or unexpected behavior.

As part of the internal review process, UW Medicine Compliance, IT Information Security, and other relevant offices will review AI use cases to assess potential privacy concerns. Additional offices, including UW Privacy, may also be consulted depending on the use case (e.g., learner data). This evaluation helps determine whether the tool is appropriate for implementation and whether potential risks related to protecting sensitive information are mitigated before moving forward.

Through the internal review process, use cases are evaluated to assess risk factors, including accuracy and reliability. UW Medicine reviews whether an AI tool performs reliably for its intended purpose, population and workflow. Internal review may include vendor evidence, experience from other organizations, local testing and controlled pilots. Because performance may vary or change over time and even mature AI tools may generate errors, users must verify outputs for accuracy and follow all required human-oversight and monitoring controls.

Low-risk use cases (as defined in the UW Medicine AI Policy) are not internally reviewed but must adhere to other existing UW Medicine policies and processes (e.g., IS Risk Management Standard, Privacy, Confidentiality, & Information Security Agreement (PCISA)). Individuals are responsible for reviewing AI-generated outputs before relying on or acting on them for helping identify and report potential risks, errors, or unexpected behavior.

To streamline operations and minimize duplicative efforts, UW Medicine has augmented its IT intake process to kick off internal review of an AI use case. Once an elevated risk use case (as defined in the AI Policy) is submitted through intake, the review process typically extends the standard IT timeline by 3–10 weeks, depending on factors such as the number of required risk consultations, the prioritization and complexity of the use case and, if applicable, the complexity of the vendor contract.

Use cases that have higher levels of elevated risk, including those that impact learners or the learning environment, and/or if appropriate controls cannot be implemented to mitigate risks, may require additional time to allow for a more comprehensive evaluation. These timelines reflect UW Medicine’s commitment to conducting thorough and responsible reviews.

We will continue to evaluate and refine the review approach to streamline the process wherever possible.

The internal review committee responsible for assessing and approving your AI use case depends on the level of risk associated with it. All AI use cases that require review under the UW Medicine AI Policy are reviewed initially by the AI Use Case Review Council. If the use case involves certain characteristics associated with possible elevated risk and/or if appropriate controls cannot be implemented to mitigate risks is it may require additional review and/or approval from the AI Operations Subcommittee or AI Education Review Workgroup before the use case can be piloted or further implemented.

For certain activities, learners must also receive authorization from their program, supervisor, or instructor prior to using AI tools, including UW Medicine Approved Tools or AI uses otherwise considered low risk. For more information on learner authorization requirements, see the Education Review – For Learners section in these FAQ below.

To learn about how to submit a use case for review, visit the Innovation Core AI Intake page.

Yes, if a use case isn’t approved by the Use Case Review Council, you can ask to have it escalated to the AI Operations Subcommittee, which is the next level of AI governance.

For learners, escalation should proceed through the standard program policies and procedures established for other academic appeals.

For the list of risks that will be reviewed, please see Section B.1: AI Characteristics Associated with Possible Elevated Risk of the AI Policy. Additional risks outside of Section B.1 may be identified and considered by the reviewing bodies, if deemed applicable.

In most cases, AI tools that were already in use before the policy took effect do not need to undergo internal review and may continue to be used. However, some existing tools do require internal review under the policy. See below for when AI uses need review even though they are already in use:

For clinical care and UW Medicine business operations:

1) If the tool includes generative AI (GenAI) capabilities and was already in use before September 1, 2025.

2) If new AI functionality is added to an existing tool and that functionality meets the AI policy’s internal review criteria.

For education:

1) If the tool includes GenAI capability and was already in use before September 1, 2026.

2) If new functionality is added to an existing tool and that functionality meets the AI policy’s internal review criteria.

Monitoring the ongoing use, effectiveness and risks of AI tools is a shared responsibility. The AI Technical Monitoring Council, established as part of the UW Medicine AI governance structure, helps guide and advise on how to monitor the technical components and performance of AI tools.

Business owners are responsible for overseeing how the tool is used in practice, including monitoring its effectiveness and ensuring it continues to meet operational needs. IT partners also provide ongoing technical support, help manage system performance and integrations, and assist in identifying and addressing issues that may arise over time.

In addition, individuals using AI tools are responsible for reviewing AI-generated outputs before relying on them and for helping identify and report potential risks, errors, or unexpected behavior.

This policy does not apply to all research using AI. It only applies to research studies led by a UW principal investigator that require UW IRB review and use UW Medicine data or specifically recruit UW Medicine patients, staff, faculty, or learners. Research that does not involve human subjects or does not require UW IRB review is not covered by this policy.

For guidance on using AI responsibility outside the scope of the UW Medicine AI Policy, see the UW Graduate School’s Effective and Responsible Use of AI in Research or the UW IT’s Generative Artificial Intelligence General Use Guidelines.

If an AI research study meets the requirements for review, the UW Institutional Review Board (UW IRB) is responsible for reviews and approvals. For more information on when an AI review is required for a research study, please see Section C: AI Uses Requiring Internal Review for Research of the AI Policy or visit the Innovation Core AI Intake page for instructions on how to submit your research study for review.

UW IRB review will focus on accuracy, reliability, bias, equity, privacy, security, transparency and explainability as described in the HSD’s Guidance for School of Medicine Research Involving AI to help inform a risk mitigation plan. If the research involves use of AI outside of a secure UW Medicine environment, a security review is also required. This review can occur concurrently with, or prior to the IRB review; however, the IRB approval is contingent upon completion of the security review. For information about how to obtain a security review, refer to the UW Medicine Information Security Risk Management Program website.

If a study was previously approved by the UW IRB before the updated AI Policy took effect on Sept. 1, 2026, the new requirements generally do not apply.

However, if the research study adds a new use of AI, researchers are required to submit the AI supplement with their application for a study modification, and should use the Guidance for School of Medicine Research Involving AI to develop their risk mitigation plan.

Externally reviewed studies must undergo a security review if the research involves use of AI outside of a secure UW Medicine environment and involves the use of UW Medicine data or targeted enrollment of UW Medicine patients, staff, faculty or learners. Researchers will be prompted to obtain a security review for applicable studies in the Human Subject Division’s (HSD) Request External IRB Review form.

However, HSD’s Guidance for School of Medicine Research Involving AI does not apply, and the submission of the AI supplement is not required.

The timeline for the review of a research study that uses AI is dependent on whether the research requires review by the convened IRB or is eligible for expedited review. The median turnaround time ranges from 14.5 days for expedited review to 70 days for review by the convened IRB. However, as this is a newly introduced process, review times may be slightly longer initially.

It is an extremely uncommon outcome for a research study not to be approved. More often, the IRB will issue a conditional approval or a deferral requesting additional information, clarifications or revisions. However, UW policy does allow a researcher to appeal to the HSD for a formal review of a decision in recognition that there can be honest miscommunications, misunderstandings or mistakes by any of the individuals or entities involved in applying for and conducting IRB review. Details about the appeals process can be found in the SOP Appeal of IRB or HSD Determination.

Yes, learners need instructor permission to use AI tools to complete coursework, including class assignments, projects, quizzes, and presentations. Learners also need permission to use AI tools for assessments, exams and within the clinical learning environment. See Section E of the AI Policy. Learners should ask clarifying questions to understand any restrictions.

Learners may use AI tools without instructor approval for personal studying and self-education activities, such as learning about a topic, self-quizzing, or developing study notes. However, these activities must not involve UW Medicine data, as defined in the AI Policy Glossary. If a learner needs to enter UW Medicine data into an AI tool, they must use a UW Medicine-approved AI tool. As learners may rotate to different services, they should always check with their immediate supervisors before employing an AI use case in the clinical setting.

Classroom environment: Instructors should specify in their course syllabus how learners may or may not use AI tools in the completion of graded coursework and what AI attributions are required. It is the responsibility of the learner to seek clarification from their instructor.

Clinical learning environment: UW Medicine-approved AI tools may be used by learners in the Clinical learning environment following authorization from their academic units (e.g., Graduate Medical Education, Undergraduate Medical Education, Health Professional Programs, etc.), in accordance with existing School policies requiring appropriate supervision. Such authorization may be given at the academic unit level or delegated to program instructors/supervisors. Academic or training programs may also create standards for supervision that allow or restrict learners’ use of specific AI tools in the clinical learning environment. Learners should seek approval from their clinical instructor or supervisor.

Non-clinical learning environments (outside the classroom): Academic or training programs may set standards that allow or restrict learners’ AI use in degree milestones. For example, programs can specify standards for use in preliminary exams, general exams, theses, dissertations, capstones and meta projects. It is the responsibility of the learner to seek clarification and/or approval from their instructor.

UW Medicine is still working on developing an approach to AI literacy and training for learners. The goal is to balance a coordinated schoolwide approach with the need to develop AI curricula based on the needs of individual clinical and non-clinical programs and their accreditation standards. Programs are encouraged to collaborate with UW Medicine Innovation Core to ensure alignment in educational approaches within the School and with the broad UW Medicine approach to AI.

To the extent learners are also part of the UW Medicine workforce, UW Medicine is developing an enterprise-wide approach to training its workforce (see the question “Will AI literacy and compliance training be provided to us?” in the FAQ General section).

When training in learning environments outside of UW Medicine, (e.g., external affiliated laboratories, hospitals and/or clinics), learners must additionally follow the policies and/or guidance at those outside organizations.  If there is a conflict between UW Medicine and the outside organization’s AI policies, learners must seek additional direction from their program director or the designated representative for the setting.  For example, even if UW Medicine has approved an AI tool and a learner has received authorization to use the tool, learners should confirm that use of the tool is allowed if they are training in a clinical learning environment outside of UW Medicine.

In addition, when working or learning outside UW Medicine, care should be taken to align with both organizations’ data protection policies. As an example, a HIPAA-protected tool for PHI at a hospital outside of UW Medicine should not be used for UW Medicine data, and a tool approved for UW Medicine data should not be used for data from another organization.

Instructors, faculty, supervisors should evaluate inappropriate uses of AI with professional judgment, direct communication with the learner, review of the work in question and established standards for academic integrity, professional conduct, clinical documentation and supervision. Due to concerns about bias and false positive rates, learner data should never be put into an external AI checker, even if there is a subscription and the vendor promises not to share the data. Escalation of suspected inappropriate use of AI by a learner should be done through existing academic, professional and clinical supervision processes.

Any suspected violations involving students should follow the School’s standard Student Conduct processes through the Community Standards and Student Conduct office. Concerns involving residents, fellows, postdoctoral scholars, or other learners should be addressed through the processes outlined in the UW Medicine Policy on Professional Conduct and any applicable training program policies.

When AI tools are used in instruction, assessment or grading, instructors, faculty and supervisors must ensure the tool and use case have been approved by UW Medicine for handling the relevant data, including FERPA-regulated information, and must comply with applicable academic record retention policies. Transparency regarding the use of AI in instructional design, learner assessment or evaluation is strongly encouraged.

AI-checking tools were removed from UW learning management systems due to concerns with bias and false positive rates. Entering learner data into external tools is a violation of FERPA. Learner data should never be put into an external AI checker, even if you have a subscription and the developer promises not to share the data. Escalation of suspected inappropriate use of AI by a learner should be done through existing academic, professional and clinical supervision processes.

This policy does not apply to UW School of Medicine faculty acting in a clinical role at Fred Hutch, Seattle Children’s, WWAMI clinical sites or other non-UW Medicine site. It does apply when UW School of Medicine faculty, including those throughout WWAMI, are acting in a School of Medicine educational role or in any other UW Medicine capacity. For learners, please see the question around policies to follow when working in settings outside of UW Medicine.

State law, as well as School of Medicine educational program policies, require appropriate oversight and supervision of learners in learning environments, including learners’ use of AI tools. In learning environments, AI use must occur under appropriate supervision and in alignment with existing school and training program policies. Faculty, instructors and supervisors, including trainees in a teaching capacity, play a critical role in authorizing, supervising, and assessing AI use in ways that preserve the development of independent human judgment, critical thinking, communication skills, professionalism and compassionate patient care. While individuals remain responsible for the accuracy and integrity of their own work and documentation, educators are responsible for guiding learners in the appropriate and ethical use of AI and for ensuring that technology supports — rather than replaces — the human knowledge, mentorship, and professional growth essential to healthcare and biomedical education.

Stating your expectations in both the syllabus and assignment instructions helps students mitigate unauthorized use of AI. The University of Washington (UW) has advice and sample course-level and assignment-level policies that you can adapt to your needs. However, there are a few additional UW Medicine considerations for UW School of Medicine instructors as you write your AI course-level and assignment-level AI policies:

• UW Medicine AI Policy: While UW does not have a university-wide AI policy, UW Medicine does have an AI policy that covers education. All uses of AI must align with UW Medicine’s AI policy and other applicable UW and UW Medicine policies.
• UW Medicine data: UW Medicine data can only be used with AI tools approved by UW Medicine. At this time, UW Medicine Chat is the only AI chat tool approved for use with UW Medicine data. Purple, UW’s AI platform has not been approved for use with UW Medicine data, including protected health data (PHI) and other sensitive data. However, Purple can be used with coursework that does not include UW Medicine data and other publicly available data. For more on how UW Medicine can use Purple, please see the addition information on Purple in this FAQ.

You can subscribe to the bi-monthly Innovation Core Newsletter (“The Roundup”) for the latest AI and innovation-related news or visit the UW Medicine Innovation Core website.

For questions not answered by the AI Policy or this FAQ, please contact UWMInnovationCore@uw.edu.