Artificial intelligence/machine learning in Alaska Native healthcare systems: symposium perspectives Authors: Vanessa Hiratsuka, Angelica Pritchard, Gretchen Day, Timothy Thomas, Kyle Wark, Aliassa Shane, Robert Onders, Tina Hernandez-Boussard, Brian Rice Published: 2026 ABSTRACT ABSTRACTArtificial intelligence and machine learning (AI/ML) utilisation in Alaska Native and American Indian (AN/AI) healthcare presents an opportunity to address persistent health disparities and improve the well-being of Alaska Native communities. The 2025 Artificial Intelligence in Alaska Native Health Care Systems symposium occurred in Anchorage, Alaska, with 31 individuals participating in facilitated conversations between various stakeholders and specialists at the Alaska Native Tribal Health Consortium, Southcentral Foundation, Maniilaq Association and Stanford University. The overarching goal of the symposium was to bring together partners within the Alaska Tribal Health System (ATHS) to promote: (1) AN/AI peoples’ guidance in support of Tribal sovereignty for AI/ML implementation within the ATHS, (2) expanded technical capacity and (3) centralised guidance and expertise to lead future AI/ML research and implementation with community-defined benefits. During the event, small group discussions focused on the following topics: (1) common AI/ML applications, (2) concerns surrounding AI/ML, (3) current or potential policies guiding AI/ML use and (4) how to build trust in AI/ML models. Themes discussed were data democratisation, AI/ML development and use needs to occur under the principles of data sovereignty, and AI/ML quality improvement and research are needed and distinct. The symposium was an important step in considering how to use AI/ML to optimise AN/AI health and well-being through collaborative partnerships and services. Introduction There is considerable and growing interest in the use of artificial intelligence and machine learning (AI/ML) in rural and Indigenous healthcare [1–8]. AI/ML have been used by health systems to help with administrative needs such as workflow optimisation, electronic health record (EHR) charting and telemedicine patient monitoring. In imaging studies, AI/ML can help interpret biopsies, endoscopic exams, magnetic resonance imaging studies and ultrasound results. They can aid in the diagnosis of cardiovascular issues, infectious diseases, cancer, ophthalmological conditions and even psychiatric issues. With regard to pharmaceuticals, AI/ML are utilised in drug discovery and in identifying drug‒drug interactions. AI/ML are even being used to develop therapeutic chatbots to help with psychological conditions [9]. Guidelines for the use of AI/ML in healthcare have been issued by the World Health Organization, National Academy of Medicine [10], the US Food and Drug Administration and the Centers for Disease Control & Prevention (CDC) [11]. However, there are no specific guidelines or ethical considerations provided for the implementation of AI/ML use within the Alaska Native and American Indian (AN/AI) tribal healthcare systems. Later in this paper, themes for guideline recommendations for the implementation of AI/ML for research or service delivery highlighted in the 2025 symposia will be elaborated on, but to understand the richness of the dialogue, it is important to understand the factors that shaped the dialogue held that day. The deployment of AI in healthcare, particularly in rural and Indigenous communities, requires careful consideration of cultural and contextual factors [12]. AI-driven technologies must be adapted to meet the specific needs of these populations, ensuring that they do not exacerbate existing health disparities but instead contribute to their reduction. Responsible AI (or ethical AI, or trustworthy AI) is not a way to give technology the responsibility for its actions and decisions; rather, it is a planful development and use of AI/ML that requires more responsibility and more accountability from the people and organisations involved [13]. Responsible AI accounts for the decisions and actions of AI applications, and for their own decision of using AI in a given application context. Structures of power, participation, and access to technology determine who can influence which decisions or actions are being automated, which data, knowledge, and resources are used to learn from, and how interactions between those who decide and those who are impacted are defined and maintained. Responsible AI is more than the ticking of a predetermined grouping of ethical “boxes” or the development of some set of add-on features in AI/ML systems [14]. The Alaska Tribal Health System The Alaska Tribal Health System (ATHS) is a statewide voluntary affiliation of Tribal Health Organizations (THOs) providing health services to nearly 180,000 AN/AI people. Each of the Tribal organisations is part of the Alaska Tribal Health Compact, an umbrella agreement that sets forth the terms and conditions of government-to-government relations between the 229 Federally-recognised Tribes in the state of Alaska and the U.S. Federal government. The compact authorises Tribes and Alaska Native Health Organizations to operate health and health-related programs. Each THO retains autonomy with regard to health priorities, services, and policies in its respective service area. THOs administer village clinics, sub-regional clinics, regional hospitals, dental, behavioural/mental health, substance abuse counselling and treatment, health promotion and disease prevention programs and environmental health services in twelve regions throughout the state. Each community served by a THO is a federally recognised Tribe. Each Tribe’s governing council has authorised its THO to act on its behalf in negotiating with the Federal Indian Health Service (IHS) to provide healthcare services under Title III of the Indian Self-Determination and Education Assistance Act of 1975. The Alaska Native Tribal Health Consortium (ANTHC), located in Anchorage, is the heart of the ATHS. In conjunction with the Southcentral Foundation (SCF) – the THO serving southcentral Alaska, the sole tertiary care hospital in the ATHS. ANTHC is led by a Board of 15 AN/AI leaders from around the state who represent its customers and provide statewide services in speciality medical care. THOs regularly share EHR data with SCF; the majority of THOs use Oracle Health EHR (previously named Cerner). Features of Alaska Native healthcare – data sovereignty Indigenous data sovereignty is the right of a Federally recognised Tribe and the THO’s operating under the authority of a Tribe or Tribe(s) to govern the collection, ownership, and use of its own data [15]. To date, there has not been a policy governing the use of AI/ML in ATHS and service delivery formally implemented, thus making these efforts to formalise policies and procedures for integrating the use of AI/ML important. Governance of health research in the ATHS occurs through community research review policies and procedures that are focused on promoting community benefits and sociocultural relevance of research, minimising stigma and harm, and maintaining data ownership and control [16,17]. The following is an example of the rigorous research review process used at [organisation blinded], the largest, most comprehensive THO in the United States, and Alaska’s second-largest health employer offering an array of health services statewide. This example illustrates the multiple layers of policy and oversight that must be considered when implementing the use of AI/ML for research within the ATHS. Prior to the initiation of a study, all research conducted by or affiliated with [organisation blinded] must be reviewed and approved by both the IHS Alaska Area Institutional Review Board (AAIRB) and the ANTHC Health Research Review Committee. Research conducted with a THO must be approved by that THO as well, prior to initiation. The AAIRB is constituted of healthcare providers, researchers and community members. Over half of the AAIRB members are AN/AI people. The ANTHC Health Research Review Committee, a Board Committee, oversees health research conducted within [blinded] as well as research involving [blinded] staff and facilities throughout the state. Tribal approval is sought to affirm that research is conducted to benefit AN/AI peoples, assuring that Alaska Native communities are not harmed, that proposed research studies address health needs pertinent to AN/AI peoples, and that results are disseminated back to AN/AI peoples. Prior to HRRC review, all research abstracts, manuscripts, and proposals are internally reviewed by the Abstracts, Manuscripts and Proposals Review Committee (AMP-RC), which provides feedback in the form of recommendations to authors and investigators. AI/ML in AN healthcare systems will need to mirror the values of the AN/AI peoples, and the healthcare system support. Just as the ATHS has been built to reflect and uphold the AN values of Alaska Tribes [18,19], AI/ML use within this unique system must carefully consider the complex context of a geographically rural/remote, minimal infrastructure environment and link cultural and community values. The overarching design will need to be comprehensive enough to adhere to identified AN values, share user control and/or regulation, and link values to formal system requirements and concrete functionalities, while maintaining low-cost implementation and maintenance, as the ATHS, like the Indian Health Service it is part of, is chronically underfunded [20]. Features of Alaska Native healthcare – rurality With the largest landmass and the lowest population density of any state within the US, the rurality of Alaska is one of its defining features. Over 80% of Alaska communities are not connected to the road system; most of these are AN communities with a population of 200 people or less. Transportation is primarily by small aircraft year-round, by boat, or by snowmobile, depending on the season. In response to the challenge of providing high-quality healthcare without a road system, the ATHS developed the unique “hub and spoke” model described previously [21]. This model employs the network of THOs who administer “spoke” village clinics staffed primarily by non-physician Community Health Aide/Practitioners (CHA/P), and regional “hub” hospitals throughout rural Alaska. Patients living in remote villages can receive primary or routine unscheduled care from the CHA/P, but for any medical issue requiring a physician-level of care, they need to travel to the hub hospital, which has an emergency department, inpatient beds and physician services. Any patient with emergency health needs requires an air medical ambulance (or medevac) because ground transport or ground ambulances are not an option off the road system, so the rates of medevacs are at least 40 times higher than anywhere else in the US [22]. Medevac decision-making is extraordinarily complex because of the minimal resources to evaluate a patient available in the villages (i.e. no laboratory or medical imaging), the broad range of patient emergencies that may present, the need to coordinate evaluation and care remotely between a physician at the hub hospital and the CHA/P in the village, as well as the non-clinical issues such as weather and time of day. In addition, medevacs represent a limited resource, often with a single crew covering a huge service area, a high financial cost and a high crash risk [23], yet they are needed daily in a region where extreme rurality leads to excess mortality for emergency medical conditions [24]. Given the complexity and the high-risk and high-reward nature of medevac services, there is potential for AI/ML to support clinical decision-making and improve the safety and quality of medevac utilisation. The use of AI/ML to support clinical decision-making for medevac utilisation was the primary example used throughout the 2025 symposia. Past efforts In 2022, prior to the 2025 symposia, the AIM AHEAD-funded joint effort by ANTHC Research Services and the Center for Alaska Native Health Research at the University of Alaska Fairbanks worked to: (1) determine the use of AI/ML in health research and healthcare involving AN people throughout Alaska, (2) introduce participants to AI/ML methods and applications in health research and healthcare and (3) develop research and clinical opportunities for AN health AI/ML applications. This project included an online survey that assessed baseline AI/ML knowledge and attitudes among clinicians, researchers, community members and AN leadership. The survey was followed by a series of eight one-hour educational live Zoom presentations conducted by nationally acknowledged leaders in the field and concluded with a symposium called Artificial Intelligence/Machine Learning Prospects for American Indian/Alaska Native Health held in September 2022, in Anchorage, Alaska. The participants at the symposium were given the opportunity to submit draft AI/ML research proposals to be workshopped at the event. All events were free and promoted to ATHS staff and the University of Alaska (UA) faculty/research staff and students. Of the eighty-one individuals who responded to the baseline survey, 30% were [blinded] researchers or faculty members, 30% were ATHS researchers or program staff, 13% were ATHS clinicians, 13% UA students, and the remainder was a mix of state and federal public health workers, ATHS leadership, and AN community members. More than a third (35%) of these respondents said they had never heard the term AI/ML, or if they had heard it, they knew nothing about it. Only 12% of the respondents indicated having ever used AI/ML for research projects or clinical care. In addition, almost half of the survey respondents (48%) said they believed AI/ML would either not improve health outcomes for AN peoples, or they just didn’t know. At the 2022 Artificial Intelligence/Machine Learning Prospects for American Indian/Alaska Native Health symposium, 31 persons attended, and more than half (58%) brought draft AI/ML health research proposals to be workshopped. Project organisers were optimistic that the use of AI/ML in health research and healthcare would bloom throughout the ATHS. However, that did not happen, which led to the 2025 symposia to continue the discussion around the use of AI/ML in research with tribal communities. Consortium for Alaska Native artificial intelligence development (CANAID) The 2025 CANAID project was a second attempt to build off of prior efforts and create a system to incorporate AI/ML applications into AN health research and healthcare. A team of researchers from ANTHC Research Services and Stanford University began research into using AI for medevac utilisation concurrently, with the development of the 2022 AIM-AHEAD-funded project to assess and improve the understanding and use of AI/ML in AI/AN health research and healthcare in Alaska. As no precedent existed within the ATHS for this kind of work, a second line of research was funded by AIM-AHEAD in 2024 to look more closely at the implementation of AI/ML within the ATHS. Symposium design The overarching goal of the symposium was to bring together partners within the ATHS to promote: (1) AN/AI guidance in support of Tribal sovereignty for AI/ML implementation within the ATHS, (2) expanded technical capacity and (3) centralised guidance and expertise to lead future AI/ML research and implementation with community-defined benefits. To accomplish this, a snowball sampling technique was employed [25]. Because the team already included membership from each partner organisation to be included in the symposium, we used those as starting points and asked within each organisation who should be involved in a conversation about AI/ML within the health system. As each respondent identified a new viewpoint to include in addition to our starting point of technical, ethical and research (e.g. legal, compliance and security), we returned to the other partner organisations and asked them to identify counterparts to include. This iterative process was performed over two months, and then three additional months were spent to find a suitable date that could work to allow maximal in-person attendance at a site in Anchorage. Travel funds were offered to interested participants. Many of the people who attended the 2022 symposium also participated in the May 2025 CANAID Symposium; however, at this event, we had more robust representation from clinical, research, IT/programming, compliance and leadership departments at each of the participating ATHS organisations. Of the 31 individuals in attendance, 27% identified as researchers, 24% were from IT/programming, 17% held leadership roles, 12% were clinicians, 5% were in compliance roles, and 15% were in other groupings. In addition, there was representation as follows from the different partnering institutions: ANTHC (38%), SCF (27%), Maniilaq Association (15%), Stanford University (8%), and other organisations, including CDC- Arctic Investigation Program (8%) and Other (4%). Informal analysis of survey results suggests that overall knowledge, use, and awareness of AI/ML have increased substantially since the 2022 symposium. In addition, everyone attending expressed confidence in the ability of AI/ML to benefit health outcomes. The 2025 symposium facilitated a conversation between various stakeholders and specialists at ANTHC, SCF, Maniilaq Association and Stanford University. During a portion of the event (Table 1. Symposium Agenda), small group discussions focused on the following topics: (1) common AI/ML applications, (2) concerns surrounding AI/ML, (3) current or potential policies guiding AI/ML use and (4) how to build trust in AI/ML models. These were intended to allow opportunities for engagement between personnel and/or organisations that might not otherwise have the framework to share ideas on such topics. Table 1.Symposium agenda.TopicPanelistBlessing Welcome Overview of CANAID project Future Work for AIM AHEADGreg NothsteinProgram Manager, Alaska Native Tribal Health Consortium Timothy Thomas, MDDirector, Clinical and Research Services, Alaska Native Tribal Health Consortium Robert Onders, MD, JDMedical Director, Maniilaq Association Usha Sambamoorth, PhD, MAProfessor of Pharmacotherapy, Associate Dean for Health Outcomes ResearchWhat is Artificial Intelligence for healthcareTina Hernandez-Boussard, PhD, MPHAssociate Dean of Research and Professor of Medicine, Biomedical Data Sciences, Surgery and Epidemiology & Population Health, Stanford UniversityWhat frameworks are used to make AI in healthcare ethical and equitable?Vanessa Hiratsuka, PhD, MPHSenior Researcher, Southcentral FoundationHow are data efforts approved within the ATHSTimothy Thomas, MDDirector, Clinical and Research Services, Alaska Native Tribal Health ConsortiumHow is AI used currently in healthcare in Alaska?Julie Beans, MPHResearcher, Southcentral Foundation Ben KelshDirector of Clinical Informatics, Maniilaq Association Kyle Wark, MAResearcher, Southcentral FoundationTechnical nuts and bolts of AI next stepsTodd Leroux, PhDSenior Director of Analytics, Alaska Native Tribal Health Consortium Eric Wiener, MSData Infrastructure Manager, Alaska Native Tribal Health ConsortiumWhat are our current viewpoints surrounding AI in healthcare and research?Gretchen Day, MPHStatistician, Alaska Native Tribal Health Consortium Angelica Pritchard, BSSenior Research Coordinator, Stanford UniversityCase studiesBrian Rice, MDCMClinical Associate Professor, Department of Emergency Medicine, Stanford UniversityMoving ForwardRobert Onders, MD, JDMedical Director, Maniilaq Association Symposium agenda. The expanse and presence of AI/ML in society’s experiences were made apparent during one of the CANAID symposium’s discussion groups. For instance, current practices and applications with AI/ML include a range of medically related tasks, such as dictation, biowriting and service utilisation. In addition, other uses were referenced, such as landscape design, maps and knowledge finding. These all build to showcase how AI/ML is beginning to change our everyday lives. The participants also proceeded to share potential concerns regarding AI/ML for Alaska community members. Individuals showcased the possibility of communication breakdowns and the concerns surrounding interpretation. However, the main concerns revolved specifically around topics such as privacy, accuracy and trust. These findings indicate the need for guardrails and transparency as models are built and established in regular applications. Interestingly, when we engaged with policies and procedures, it was discovered that some of the stakeholders present had recently approved an organisational policy on the use of generative AI scribes in primary care visits. Other participants spoke of the desirability for consent, standardisation, and governance to be included with policy-making and best practices. Finally, for best practices, symposium attendees shared how trust could be built with AI/ML. A common insight was to engage a broad range of interest holders from community members to administrators in the process and discussion, from initiation to execution and maintenance. There was a belief that transparency, consent, and oversight would improve how AI/ML systems could be used. In addition, the use of system metrics to verify performance and behaviour was recommended. Finally, accountability was a theme generated as a need for building trust amidst AI/ML models. Example AI/ML utilisation scenarios Three example scenarios of how AI/ML could be utilised in healthcare (Table 2. Example AI/ML utilisation scenarios) were presented in a small group discussion format to allow symposium participants an opportunity to discuss approaches to potential AI/ML activities and gain a contextual, concrete, and in-depth understanding of the barriers, concerns, opportunities, gaps and interests of stakeholders. Each scenario described the overview and problem, followed by a set of discussion questions for symposium participants to respond to in small discussion groups. One person from each group volunteered to be a notetaker. After a brief discussion, each small group presented an overview of their discussion to the whole group. Themes from the discussion of these scenarios and reflection on the overall symposium were: (1) AI/ML development and use needs to occur under the principles of data sovereignty, (2) data democratisation and (3) quality improvement and research efforts are both important and distinct. Table 2.Example AI/ML utilisation scenarios.ScenarioOverview and problemDiscussion prompts1 •A THO wants to implement a sepsis early detection tool offered by Cerner at the hospital level, which has never been tested on local data•A THO wants to develop an early warning tool that identifies elderly patients living in villages at risk for emergencies using their data •Are there different pathways for approval for these? Should there be?•Can we find a common pathway that meets the needs of patients, providers, administrators and others? 2 •The predictive model is successful in predicting patients at high risk for emergencies•We can potentially provide care at the village clinic using telehealth to avoid the patient getting so sick they need to fly out •Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes?•How about becoming part of routine clinical care? 3 •Our medevac model can provide real-time information to providers about how likely the patient is to benefit from a medevac or how likely they are to deteriorate rapidly•This could become part of the evaluation of a village clinic patient to decide if they need to fly out •Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes?•How about becoming part of routine clinical care? Example AI/ML utilisation scenarios. •A THO wants to implement a sepsis early detection tool offered by Cerner at the hospital level, which has never been tested on local data A THO wants to implement a sepsis early detection tool offered by Cerner at the hospital level, which has never been tested on local data •A THO wants to develop an early warning tool that identifies elderly patients living in villages at risk for emergencies using their data A THO wants to develop an early warning tool that identifies elderly patients living in villages at risk for emergencies using their data •Are there different pathways for approval for these? Should there be? Are there different pathways for approval for these? Should there be? •Can we find a common pathway that meets the needs of patients, providers, administrators and others? Can we find a common pathway that meets the needs of patients, providers, administrators and others? •The predictive model is successful in predicting patients at high risk for emergencies The predictive model is successful in predicting patients at high risk for emergencies •We can potentially provide care at the village clinic using telehealth to avoid the patient getting so sick they need to fly out We can potentially provide care at the village clinic using telehealth to avoid the patient getting so sick they need to fly out •Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes? Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes? •How about becoming part of routine clinical care? How about becoming part of routine clinical care? •Our medevac model can provide real-time information to providers about how likely the patient is to benefit from a medevac or how likely they are to deteriorate rapidly Our medevac model can provide real-time information to providers about how likely the patient is to benefit from a medevac or how likely they are to deteriorate rapidly •This could become part of the evaluation of a village clinic patient to decide if they need to fly out This could become part of the evaluation of a village clinic patient to decide if they need to fly out •Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes? Who would need to be engaged and authorise that becoming part of a trial to see how it impacts outcomes? •How about becoming part of routine clinical care? How about becoming part of routine clinical care? AI/ML development and use need to occur under the principles of data sovereignty Participants discussed centring data sovereignty as a driving issue, as they described using concepts of the CARE principles (collective benefit, authority to control, responsibility and ethics) [26] in ATHS data use, alongside those set out by FAIR [27,28], the Common Rule (OHRP, 2017) and HIPAA. The participants referred to Tribal governance when talking about data governance. The participants reflected on how application software selection and use are driven by the priorities and values of AN peoples, from data management and analytics (e.g. tools to organise, analyse and visualise EHR datasets) to business optimisation (e.g. tools to streamline operations, project management and customer engagement). The participants had a dialogue on collaborating on intra- and interorganisational AI/ML usage that maintained data quality, security, and compliance with data sovereignty. They were interested in ongoing conversations on managing the risks associated with increased data access, ensuring data is used responsibly, and ensuring data ownership and usage fall within the priorities and values of AN peoples. Data democratisation Awasthi and George [29] described data democracy as the act of opening organisational data to as many employees as possible, given reasonable limitations on legal confidentiality and security, allowing for intra-organisation open data use within an organisation’s culture and needs [29]. THOs currently use reports to allow system administrators and healthcare providers the ability to explore the EHR system to track sub-population healthcare quality goals compared to THO and national benchmarks, and to develop action lists to support the treatment of patients with specific healthcare needs. Symposium participants discussed interest and willingness to develop and implement approaches to AI/ML that met the FAIR (Findable, Accessible, Interoperable and Reusable) principles [27,28] through collective empowerment of data consumers and data providers and development/sharing of tools (i.e. policies, code, experimentation environments, reports). Participants described interest and progress in developing self-service analytics tools within their THO’s EHR, training more clinical and administrative leadership on data and analytics, cross-THO collaboration and knowledge sharing, and promotion of data value. Internal policies are already being developed and were being shared during the symposium. The participants pointed out how great it was to have subject matter experts working within THOs in one place and how some partners had a broader range of experts/participants than others, as AI/ML use in Tribal healthcare has already begun in two of the THO partner organisations [24,30]. The participants emphasised the importance of empowering THO employees with actionable data, allowing THOs to make faster, more informed decisions and improve clinical and operational efficiency. Quality improvement and research efforts are both important and distinct The participants discussed the implementation of AI/ML tools in the ATHS primarily in the form of quality improvement efforts. The participants expressed interest in AI/ML implementation for diagnosis and treatment and operations (e.g. operational efficiency, risk assessment, predictive analytics, quality indicator reporting, and staff training and education). Research was most frequently described in the development of AI/ML, particularly in testing use cases and validating algorithms with AN/AI EHR data to help address the systematic under-representation of AN/AI patients in AI/ML research datasets. Operational efficiencies were frequently discussed as quality improvement, ensuring compliance with regulatory standards and accreditation and in monitoring and reporting quality metrics. Quality improvement was posited to occur in a variety of operational areas, including uncovering patterns and trends across clinical providers related to treatment, discovery of time-consuming administrative tasks, and summarisation and drafting of documentation through the use of AI/ML. The participants also noted the need for processes to confirm that AI/ML tools are operating as intended through ongoing quality assurance and quality control checks. Conclusion Within a span of three years, stakeholders’ AI/ML awareness, knowledge, and attitudes had changed drastically. The 2025 AI Symposium targeted a homogenous group of stakeholders – individuals working with ATHS EHR data – allowing for focused dialogue on AI/ML within a similar context, yielding direct outcomes and collaborations during the symposium, unlike the 2022 event. To increase feedback and discussion, we included small group discussions and case studies at multiple points in the agenda rather than at the end of the day. Discussion was facilitated by prompts related to didactic presentations to seed the conversation and allow participants opportunities to get to know one another, share ideas, and learn about THO-specific AI/ML issues. AI/ML areas of use mentioned by participants focused on operational support to enhance THO efficiency at the individual level, with interest in applications to increase operational efficiency at other ecological levels, such as autonomous resource distribution on a regional level. Several meaningful and tangible actions occurred during the symposium. First, an ANTHC stakeholder described and shared the statewide organisation’s AI use policy. Second, a rural THO described their recent implementation of Cerner AI Scribe in primary care visits. Third, the participants shared their interest in collaborating further on AI/ML, using an existing, regular meeting called “Shared Tribal Analytics and Reporting (STAR)”. Finally, partnerships were discussed for future AI/ML research projects related to prevention and interventions in THO health priority areas (e.g. cardiovascular disease). Our symposium was an important step in considering how to use AI/ML to optimise AN/AI health and well-being through collaborative partnerships and services.