Definitions
- Caregiving
-
This challenge uses the definition of caregiving from the National Strategy to Support Family Caregivers. “Caregiving” includes a wide variety of tasks to meet the individual needs of the person receiving support. Level of effort can vary significantly — from assistance with one (or a few) specific tasks to live-in assistance that includes medical tasks. It most often describes the support provided to help people live in their own homes and communities, rather than in nursing homes and other institutions. The level of support can vary based on the care recipient’s needs and preferences. Examples include assistance with personal care and hygiene, medication and symptom management, housekeeping, transportation, recreational and social activities, and coordination of care across multiple health and social providers.
- Family caregivers
-
This challenge uses the definition of family caregivers from the National Strategy to Support Family Caregivers. “Family caregivers” include people of all ages, from youth to grandparents; people with and without disabilities; people providing care from a distance; and people meeting a wide variety of needs, such as supporting people with intellectual and developmental disabilities (I/DD) across the lifespan, caring for people with serious or progressive illnesses like dementia and cancer, and assisting with daily tasks that can be challenging for older adults and people with disabilities. The term also recognizes that a single person may receive care from multiple family caregivers. Some family caregivers are part of the “sandwich generation,” raising their own children while also caring for aging parents.Family caregivers often assist with daily tasks such as bathing, feeding, errands, medication reminders, transportation, and companionship. Many are also providing care that goes beyond basic help and involves ongoing medical tasks and health-related decision-making and monitoring at home. While family caregivers typically do not receive formal training, they assume caregiving responsibilities to support their loved one’s preferences to stay in the home environment and in the community.
- Direct care workers
-
Direct care workers are paid professionals with various job titles and roles, including personal care attendants, home health aides, nursing assistants, direct support professionals, job coaches, and more. Their services can range from assisting with personal care and activities of daily living to providing support that is essential to someone’s ability to live, work, socialize, volunteer, and participate in their community.The American Network of Community Options and Resources (ANCOR) State of America’s Direct Support Workforce Crisis 2025report shows that severe staffing shortages among community-based disability service providers persist nationwide. These are hindering service access, forcing referral refusals, delaying or cutting programs, and undermining quality of care. Despite slight improvements from 2024, there is still a need for stronger recruitment, retention, and policy action to stabilize the direct support workforce.
- Artificial intelligence
-
For this prize competition, artificial intelligence(AI) is defined as a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. AI systems use machine and human-based inputs to perceive real and virtual environments, abstract such perceptions into models through analysis in an automated manner, and use model inference to formulate options for information or action.
Frequently Asked Questions
Webinar Materials, Recording, and Slides
- Will the recording and presentation be available afterward?
-
Yes, they can be found at the bottom of this page, in the events section.
Tracks and Applications
- How many PIs can be included in the application?
-
There is no set or recommended number of principal investigators (PIs). There must be only one designated primary team leader or entity for each team to serve as the main point of contact for the application and receive the prize money.
- Can applicants submit to both Track 1 and Track 2?
-
Yes, multiple entries are permitted. Each solution must be sufficiently novel and unique per Track, and not be a minor modification of a prior submission.
- Can your solution be in both tracks at the same time?
-
No, submit your application to the Track that best fits your solution. Consider what is the main purpose or primary use case for your solution and which track does it best align with.
- If submitting two entries, can the submissions be staggered, or do both need to be submitted at the same time?
-
Applications can be submitted separately, but both must be submitted by the deadline.
- If the user of the solution is paid caregivers but enablers are home agencies, what track would the solution best align with?
-
If the main user of the AI tool is the paid caregiver while working directly with the client, then it would be Track 1. If the solution is a more comprehensive tool to support and manage the direct care workforce, it would fit in Track 2.
Caregiver Definition and Scope
- How does ACL define “caregiver” for this challenge?
-
The definition of “caregiver” is provided on the Challenge Definitions section.
“This challenge uses the definition of caregiving from the National Strategy to Support Family Caregivers. “Caregiving” includes a wide variety of tasks to meet the individual needs of the person receiving support. Level of effort can vary significantly — from assistance with one (or a few) specific tasks to live-in assistance that includes medical tasks. It most often describes the support provided to help people live in their own homes and communities, rather than in nursing homes and other institutions. The level of support can vary based on the care recipient’s needs and preferences. Examples include assistance with personal care and hygiene, medication and symptom management, housekeeping, transportation, recreational and social activities, and coordination of care across multiple health and social providers.”
- Does your definition of community include senior living facilities and their caregiving staff?
-
No, this challenge does not include institutional care settings (e.g., assisted living facilities, memory care facilities) in its scope.
- Is in-home hospice care an eligible community?
-
Yes, in-home hospice is an eligible care community.
- For Track 2, does the focus include HCBS providers such as adult day service providers, and are employment support professionals and vocational rehabilitation staff considered “caregivers”?
-
Yes, HCBS providers that deliver individualized, community‑based day services are an eligible population of focus for the challenge. Day programming models that support people directly in the community — rather than in a centralized facility — align with the intent of the challenge.
Yes, employment support professionals and vocational rehabilitation providers are an eligible population of focus for the challenge. The expectation is that providers deliver caregiver support as part of supported and customized employment services on an individualized basis in community settings.
- Must the family caregiver live with the care recipient, or can the caregiver be remote from the care recipient?
-
Family caregivers do not need to live with the care recipient. Family caregivers can be family, or even friends or neighbors, who provide support for an individual with disabilities or older adult. They can live together or apart. You can find more information on the Challenge Definitions, FAQs, and Resources page.
- Will the challenge consider AI solutions that support caregivers of neurodivergent youth and adults, not only aging populations?
-
Yes, caregivers for adults and children with disabilities, including those who are neurodivergent, are in scope for this challenge.
- Can solutions be specific to a certain type of caregiving, for example providing care for a person recovering from stroke or providing support for someone with cancer?
-
Yes, the solutions can be tailored to certain populations or care needs.
- Are there any geographies, populations, or conditions that ACL is prioritizing for this challenge?
-
Generally, there is no preferred type of application. At least 50% of Track 1 winners will include solutions targeted at family caregivers. Additionally, solutions that focus on the areas defined in the meritorious prize criteria will be eligible for an additional award up to $50,000 on top of the $100,000 prize. You can find the relevant meritorious prize criteria on the Track 1 and Track 2 tabs under the “Prizes for Phase 1” sections.
Phases, Timeline, and Competition Structure
- Can applicants enter directly into Phase 2 or Phase 3?
-
No. You must be a winner in Phase 1 to move to Phases 2 and 3.
- If selected as a finalist/prize winner for phase 1, does that automatically qualify us for phase 2?
-
Phase 1 winners are automatically qualified for Phase 2.
- What is the prize amount for Phase 2 and Phase 3?
-
The prize amounts for Phases 2 and 3 have not yet been announced.
- Can you share an overview of what Phases 2 and 3 may look like?
-
Most details about Phases 2 and 3 have not been released to the public yet. We expect Phases 2 and 3 to each be approximately one year. We will be releasing more information about expectations for testing prior to the start of Phase 2. Early in Phase 2, we will be requesting the winning Phase 1 teams to submit a more detailed testing plan which will receive additional federal feedback and technical assistance to refine.
- How long is Phase 1?
-
Phase 1 is considered between now and the submission of Phase 1 applications on July 31, 2026. Phase 2 will start after the announcement of Phase 1 winners.
- What would be a good timeline for the implementation to make sure we are ready for Phase 2?
-
We are expecting the timeline between Phase 1 announcement of winners and Phase 2 submissions to be about a year. We plan to finalize the Phase 2 timeline after Phase 1.
Budget, Prize Use, and Funding Rules
- Please share more about any ongoing administrative and reporting requirements if awarded, the extent of government governance over winner’s business with an award, and any restrictions on how the funds can be used.
-
This opportunity is structured as a prize challenge, which is distinctly different from a grant or a contract. The prizes are awarded based on the quality of your application measured against the judging criteria. You do not need to submit a budget and there are no requirements for how the prize money can be used. There is no requirement to continue to participate in Phase 2 if a team is a winner in Phase 1, though we certainly hope winning teams will continue to submit for Phase 2. There are no reporting requirements after the award. The ACL team will be assessing and providing technical assistance to Phase 1 winners to support their application to Phase 2, but there is no requirement to participate.
- Is a budget required as part of the Phase 1 submission?
-
No. Prize challenge submissions do not need a budget.
- Can applicants use non-governmental funding sources (e.g., accelerators, grants) alongside this prize?
-
Yes. Applicants can use other funds to support their work. The Eligibility and Terms tab of the Challenge website provides information on the use of federal funds.
Partnerships, Collaboration, and Deployment
- Do you connect Phase 1 applicants with home care agencies, aging network organizations, or disability network organizations for pilot partnerships, and can applicants request introductions to ACL’s relevant partner points of contact for technical and implementation‑alignment discussions?
-
In Phase 1, ACL does not directly match applicants with home care agencies or aging or disability network organizations for pilot partnerships. However, we are able to facilitate connections where we can. ACL’s website includes searchable databases of our aging and disability networks across the U.S. and its territories. ACL is planning to offer opportunities for applicants to connect with different experts prior to the application deadline, and we will share more information as those opportunities become available. Further support will be available in Phases 2 and 3, including potential connections to aging and disability networks where it can help to strengthen the solution and user experience.
- What are the planned distribution channels and types of support for promoting the winning solution(s) across the nation and among care agencies, and do you intend to support or encourage broader deployment of the winning solutions?
-
ACL intends to promote the winning projects through our established networks and communication channels to help raise awareness and support broader of AI adoption that supports caregiving. While we do not directly deploy solutions, winning solutions will receive greater visibility through the announcements of the Challenge winners.
Examples of channels we anticipate using include:
- ACL newsletters and listservs
- Webinars and learning collaboratives hosted by ACL programs
- Conference presentations or national forums where appropriate
- Will ACL support scaling through federal programs or Medicaid Waiver Systems?
-
While we will support scaling in Phase 3, the support will not be specific to Medicaid waivers and reimbursement.
- Will applicants receive technical support for Phases 2 and 3?
-
Full details about the technical assistance that will be available in Phases 2 and 3 are not available yet. Further support will be available in Phases 2 and 3 on topics such as usability testing, interoperability, and building connections to caregiving networks. Stay tuned for more information.
Evaluation Criteria & Judging
- How is ACL addressing ethical AI design for tools?
-
See the Challenge AI Principles and the judging criteria for Track 1 and Track 2.
- What is your rubric for evaluation and selection?
-
The Challenge website provides the judging criteria for Track 1 and Track 2.
- Do you have the bandwidth for general feedback if declined? Or the scoring criteria?
-
No, we do not plan to provide feedback or the application scoring for declined applications.
Intellectual Property, Privacy, and Legal
- Who owns the IP from the proposed project?
-
This is addressed on the Eligibility and Terms tab of the Challenge website.
“As a condition for winning a cash prize in this Challenge, each participant (whether participating as a team or entity) grants to ACL an irrevocable, paid-up, royalty-free nonexclusive worldwide license to reproduce, publish, post, link to, share, and display publicly the team/entity name, title, executive summary, and plain language summary components of the submission on the web or elsewhere. Each participant will retain all other intellectual property rights in their applications, as applicable.”
- Will submitted proposals be made publicly available for viewing by the general public or other submitting participants, and who will have access to them?
-
Submitted proposals will not be made publicly available and will not be shared with other applicants or external parties. However, ACL does plan to publish summary information about the winning proposals so the public and stakeholder organizations can learn about the selected solutions.
All submissions are subject to the Freedom of Information Act (FOIA). If ACL receives a FOIA request, we will review the request in accordance with federal law, including applying all applicable exemptions to protect confidential and proprietary information.
Eligibility
- What types of organizations are eligible to apply (e.g., for-profit, nonprofit, academic, government entities)?
-
Nonprofits, for-profits, government agencies, cross-organizational partnerships, teams, and individuals are all eligible. There is no preference for organization type.
- Can public entities such as state units on aging or area agencies on aging serve as lead applicants?
-
Yes, public entities such as state units on aging are eligible to apply and can be the lead applicants.
- Are there any restrictions on institutional applicants or specific organization types?
-
No preference is given to any organization type. You can see the full Eligibility and Participation Requirements on the Eligibility and Terms tab of the Challenge website.
- Would a U.S.-based company be eligible if it is partnered with a Canadian one?
-
Primary applicants must be U.S. citizens, permanent residents, or U.S.-based entities. You can see the full Eligibility and Participation Requirements on the Eligibility and Terms tab of the Challenge website.
- If selected, will this conflict with an application for another funding source, like SBIR?
-
Generally, no, but please refer to the other funding sources for their eligibility, terms, and conditions.
AI Definition
- Please define AI as it relates to this competition.
-
For this prize competition, artificial intelligence (AI) is defined as a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. AI systems use machine and human-based inputs to perceive real and virtual environments, abstract such perceptions into models through analysis in an automated manner, and use model inference to formulate options for information or action. See the Definitions section on the Challenge website.
For this challenge, this includes predictive analytics (forecasting health risks), generative AI (summarizing clinical notes), and ambient technology (sensors or listening tools that automate documentation).
- Does AI encompass robotic process automation (RPA)?
-
Yes, but with a specific caveat.
In the context of the ACL Caregiver AI Challenge, robotic process automation (RPA) is viewed as a foundational tool that achieves the "reduction of burden" principle. However, for a Phase 1 proposal to be competitive, RPA should ideally be part of an "intelligent automation" workflow where AI provides the "brain" and RPA provides the "hands.”
While RPA is a valuable tool for reducing administrative burden, this challenge specifically seeks AI-enabled solutions. We encourage applicants to use RPA for the 'execution' of tasks, but the core value of the proposal should demonstrate AI-driven intelligence — such as pattern recognition, predictive analytics, or adaptive learning, to truly empower the caregiver.
TRL 3 Understanding
- What do you mean by technical readiness of 3? How far in progress should projects potentially already be for consideration?
-
AI-enabled tools should be at Technology Readiness Level 3 or higher to be eligible for this opportunity. The challenge provides two resources to help teams understand what that means:
- Technology readiness levels for machine learning systems
- Streamlining medical software development with CARE lifecycle and CARE agent: an AI-driven technology readiness level assessment tool
Here is how Technology Readiness Level 3 (TRL 3) translates for applicants:
- It cannot be "just an idea" (TRL 1-2): If you only have a concept on paper or a slide deck describing what the AI could do, you are not yet at TRL 3.
- It must be a "proof of concept" (TRL 3): You must have moved into active R&D. This means you have a working model or a set of successful tests that prove your "analytical predictions" are correct.
- It does not need to be "real-world ready" (TRL 4+): You don't need to have tested this in a real person's home yet (that is what Phase 2 is for). You simply need to show that the "science" of your AI works in a controlled environment.
- What is the difference between Phase 1 and Phase 2 in regard to Design versus Implementation Strategy?
-
Phase 1 is about design and strategy. We require TRL 3 because we need to know your technology is scientifically viable today. Phase 1 applications require a description of the implementation approach because we need to know you have a realistic roadmap and the engineering resources to move that science into a caregiver’s home during the testing phases. You don't need a complete app yet, but you do need a complete plan to build one.
- Can early-stage companies apply who do not have any customers yet? product demo? Stealth companies?
-
Yes. You do not need a customer base, a revenue stream, or even a public brand to apply.
- Is it expected that the proposal will have a prototype completed by the July submission?
-
No, a fully functional, user-ready prototype is not required for the July submission.
- Is there a maximum level of technology readiness or is it open to everyone after level 3?
-
We welcome mature technologies, but Phase 1 is a Design Phase. We aren't just looking for the 'most finished' product; we are looking for the 'best designed' solution for caregivers. A TRL 9 company must demonstrate how they are evolving their tool to meet the Caregiver AI Principles and meet the needs of caregivers and the agencies that manage the caregiver workforce.
More Technical Questions
- Is the solution expected to be integrated to an existing system or tech stack or is the expectation to design the system from the ground-up?
-
ACL is looking for interoperability within the home environment and with an individuals’ care team. Whether you build a new tool or integrate into an old one, your Phase 1 proposal must show how your 'Sound Engineering' allows your tool to share data securely and fit into a caregiver's existing routine without creating a new tech burden.
- For tools that handle sensitive health and care-related records, how should Phase 1 applicants document their data privacy and security posture to reviewers? Are architectural safeguards and compliance frameworks sufficient, or is formal certification expected at this stage?
-
For Phase 1 (the Design Phase) of the ACL Caregiver AI Challenge, you do not need formal security certifications (like HITRUST or SOC 2). Instead, reviewers are looking for a "Security-by-Design" narrative that proves your architecture is built to handle sensitive data from day one.
At this stage, your documentation should focus on architectural safeguards and alignment with federal frameworks.
- For Phase 1 submissions, how important is demonstrating a clear AI workflow (e.g., structured caregiver input → AI processing → organized care-planning insights) versus building a fully developed application?
-
Think of Phase 1 as the architectural phase. The judges are looking for the "blueprint" of your innovation — the logic, the data standards, and the safety protocols — rather than a finished app with a polished login screen.
A clear, documented workflow (Input → Processing → Insight) proves three things that a finished app cannot:
- Technical feasibility: It shows you understand exactly what data is needed and how the AI will handle it at TRL 3.
- Clinical/Care logic: It proves your AI isn't a "black box" and follows a sensible path to help a caregiver.
- Actionability: It demonstrates that the end result (the "insight") is actually useful and not just more "noise" for a busy caregiver.
- What additional regulations are applicable if the proposed AI solution is a medical device?
-
Depending on the intended use of the AI-driven tools developed through the ACL competition, some tools may be subject to FDA regulation as medical devices while others may not. There are several existing resources and programs offered by the FDA to aid in the determination of whether a tool meets the definition of a medical device and to obtain expert feedback on device regulatory requirements that may apply. Challenge participants are encouraged to contact the FDA with any questions about potential regulation of their tools as medical devices.
Existing AI Tools
- The Challenge Team has created this table to help clarify what types of projects are eligible.
-
Type of Solution Project Description Eligible Proprietary Pivot You own an existing AI platform and are building a new, specialized tool for this challenge. YES Technology Partnership You are a care provider partnering with a developer to significantly modify or "re-train" an AI. YES Developer Build-Out You are leveraging a commercial base model (like an LLM) to build entirely new tools* YES Off-the-Shelf Scaling You are purchasing or licensing a finished AI product without making technical changes to the AI logic to implement in your caregiving setting or network. NO *Participation requirements require the team have the right to use any copyrightable works that the submission comprises. Each team leveraging an external tool must ensure the proper agreements are in place. Visit our Participation Requirements to review the full language.
Support for Partnerships
- Should applicants already have an identified AI technology partner or tool at the time of application submission?
-
Yes, for a Phase 1 submission, you must have an identified AI technology or partner. Applications for Phase 1 should have a solution designed through Technology Readiness Level 3 (TRL 3) which includes having a specific tool that has already passed basic laboratory validation.
- Many AI ideas require access to technology. What technology do we have access to that will be providing compute and license/seats for this if our ideas include that?
-
ACL does not provide direct access to compute power or software licenses. As an applicant, you are expected to bring your own technical resources — including the servers, AI models, and software seats — to the table.
- Do you have a list of approved or recommended vendors to work with if we want to work with someone to help build the prototype?
-
No, ACL does not have a list of "approved" or "recommended" commercial vendors.
Commercialization
- Are there any limitations regarding commercializing the technology during any of the phases of the challenge?
-
No, there are no limitations on commercialization.
- Could you clarify what “affordability” means in practical terms for this challenge?
-
At this stage, we are looking for applications that are considering the end cost of the solutions they are developing to ensure that the final product will actually be able to be purchased and adopted by end users. This depends on if you are designing a solution to be purchased by agencies, provider organizations, or by informal caregivers.
Resources
- FDA Digital Health and Artificial Intelligence Glossary
- Software and Systems Engineering — Software Testing Part 11: Guidelines on the Testing of AI-Based Systems
- Technology Readiness Levels for Machine Learning Systems
- Streamlining Medical Software Development With CARE Lifecycle and CARE Agent: An AI-Driven Technology Readiness Level Assessment Tool