AI for Good Impact Awards 2026

AI for Good Impact Awards 2026

Historical reference: applications for the 2026 cycle are closed. The International Telecommunication Union (ITU) ran this global recognition program for AI solutions addressing real-world challenges. The official awards page does not announce a new cycle or a new deadline, so this entry should not be read as a live application listing.

The 2026 edition recognized work across people, planet, and prosperity. It was an awards program rather than a grant: the official prize information names an invitation to the AI for Good Global Summit, recognition and exposure, and an AI for Good Impact Award trophy. The Terms and Conditions also mention promotion through official channels and possible networking, visibility, and collaboration opportunities. No fixed cash prize is listed by ITU.

At a glance

The landing page contains a small date inconsistency: its visual timeline shows 29 March, while the FAQ and the linked Terms and Conditions identify 15 March as the closing date. The Terms and Conditions are the clearest rule for the application window: submissions closed on 15 March 2026 at 23:59 CET and late submissions were not accepted. That is the date carried in this page’s deadline field.

What the awards recognized

ITU described the Impact Awards as recognition for innovative AI solutions delivering real-world impact. The program asked applicants to connect an AI system to a specific global challenge and show more than technical novelty. Projects were expected to demonstrate innovation, sustainability, and impact in their design and development, while using AI as a core component.

The three categories gave applicants a practical way to frame that evidence.

AI for People

AI for People covered human well-being, inclusion, and digital participation. ITU’s examples included healthcare and diagnosis, accessibility and assistive technologies, education and lifelong learning, language inclusion, teacher support, digital equality, support for vulnerable populations, poverty reduction, and food security.

An application in this category needed to make the human outcome clear. A description of a model was not enough by itself. A stronger account would identify who used the solution, what obstacle they faced, how the system changed the service or experience, and what evidence supports that change. For a health tool, that might mean explaining how clinicians used it and how performance was checked. For an education tool, it might mean describing the learners, the learning setting, and the measured result. The relevant proof depended on the project, but the connection between AI and human benefit had to be explicit.

AI for Planet

AI for Planet focused on environmental sustainability and climate-related action. The official examples included conservation, agricultural optimization, resource access, and other work that helps protect ecosystems or use resources more responsibly.

Applicants needed to link the system to an environmental outcome rather than relying on a broad sustainability claim. Useful evidence could include changes in resource use, the quality or timeliness of monitoring, conservation decisions supported by the tool, or the people and ecosystems reached. The important point was not a particular metric; it was a credible explanation of what changed and how the applicant knew.

AI for Prosperity

AI for Prosperity addressed economic resilience, inclusive access, public services, policy, and related systems. ITU listed focus areas such as financial systems and inclusive economic access, public-sector innovation and regulatory systems, data-driven decision-making, policy intelligence, crisis response, humanitarian systems, and peacebuilding.

This category could fit a project improving access to finance, strengthening a public service, informing policy, or helping institutions respond to a crisis. Because these systems can affect people’s rights and opportunities, the application also needed to explain safeguards, accountability, and the role of people in consequential decisions. A faster process is not automatically a better one if the application cannot explain error handling, review, and responsibility.

Eligibility and project maturity

The 2026 call was open globally. ITU listed small and medium enterprises, large corporations, non-profit organizations, academic and research institutions, government bodies, and individual contributors. The Terms and Conditions added that applicants had to be at least 18 years old at the time of submission. The program was not restricted to one country or region.

Startups and early-stage companies were explicitly excluded from the Impact Awards. ITU directed those applicants to the AI for Good Innovation Factory startup competition instead. That distinction matters: a young company should not treat the Impact Awards as a general startup prize merely because its product uses AI for social benefit.

The project itself did not have to be a finished, large-scale deployment. ITU’s FAQ said that an MVP could apply if the applicant could explain its potential impact, provide initial results, describe the innovation, and show a clear path to scaling. The page also described eligible projects as existing solutions with proven impact, while acknowledging that an MVP with results and scaling potential could qualify. In practical terms, the strongest historical applicants were likely to be teams that could show a working intervention and evidence from use, a pilot, or an evaluation—not only a concept deck.

Collaborative projects were welcome. One organization or individual had to act as the primary applicant throughout the awards process, while collaborators could be credited in the submission and recognized if the project became a finalist or winner. The physical trophy would go to the entity that submitted the application.

What the application required

The 2026 application had to be submitted through the official online platform. Each solution could be entered in one award category only; duplicate submissions of the same project across categories could lead to disqualification. The form and all submissions had to be in English.

According to the official FAQ and Terms and Conditions, the submission package included:

• A completed application form with a detailed description of the project.
• Evidence of the solution’s measured impact.
• Supporting documentation.
• Optional multimedia materials, such as a pitch deck or demo video.

The official FAQ also said that applicants without a demo video or documents could still explain the current solution and its impact in the form, and that a link to a working MVP was useful. That did not remove the need for evidence; it meant that optional media were not the only way to communicate the work.

A sensible historical application structure would have been:

1. Name the category and define the problem in terms that a non-specialist can understand.
2. Explain what the AI does, what part of the service or workflow it changes, and who uses it.
3. Show the measured result, including the baseline, population or setting, time period, and measurement method.
4. Describe sustainability, responsible use, privacy, accountability, and the path to wider use.
5. Attach only documentation that supports the claims and identify collaborators clearly.

The Terms and Conditions required information to be true, accurate, and complete. They also allowed ITU to verify information and disqualify submissions for incomplete material, false information, or failure to meet the rules. That makes careful claims more valuable than impressive but unsupported projections.

How judging was described

The official judging criteria were innovation, impact, sustainability, contribution to global challenges, and ethical considerations. The FAQ added scalability and relevance to global challenges, and said that a weighted scoring rubric would be used. A panel of experts from ITU and external partners evaluated submissions.

For applicants, those criteria translate into five questions:

• Is AI central to the solution, and is there a meaningful reason to use it?
• What changed for people, the environment, or economic and public systems?
• How was that change measured, and what limitations affect the evidence?
• Can the work continue responsibly, with a credible maintenance and scaling plan?
• What safeguards address privacy, transparency, safety, bias, human oversight, and accountability?

The last question was not a decorative ethics paragraph. ITU’s project requirements called for responsible, safe, and sustainable use of AI. An application involving health, public benefits, employment, education, or humanitarian decisions needed to make decision authority and human review understandable. A project with lower stakes still benefited from explaining data handling, known limitations, and how failures were detected.

Recognition, ownership, and publicity

The program offered recognition rather than a stated cash award. The official page lists an invitation to the AI for Good Global Summit 2026, recognition and exposure, and an AI for Good Impact Award trophy. The Terms and Conditions say winners would receive trophies, recognition during the summit, and promotion through official channels. They also mention possible networking, visibility, and further collaboration opportunities.

Applicants retained ownership of their submissions and associated intellectual property. By applying, however, they granted ITU the right to use submitted materials for promotional purposes with appropriate credit. The Terms and Conditions also say that personal and project data would be used by the organizer for the purposes of the awards program, subject to AI for Good’s privacy policy. Teams with confidential information would therefore have needed to decide carefully what evidence could be shared publicly.

Finalists were strongly encouraged to attend the summit, but the FAQ says attendance was not mandatory to receive an award. The official finalist coverage described nine initiatives across the three categories and reported that the winners would be announced at the Impact Awards ceremony during the AI for Good Global Summit in Geneva. The Terms and Conditions identify 8 July 2026 as the ceremony date for distributing prizes.

Current status and what to do now

There is no live application route for this cycle. ITU’s awards page currently labels the program “Applications closed.” The old form URL stored in the previous version of this entry redirects to an AI for Good Sandbox host form, which is a different program and should not be used to apply for the Impact Awards. The canonical awards page is the correct reference for the closed cycle and for any future announcement.

No next Impact Awards cycle or new deadline is announced on the official page. Organizations interested in a future round should monitor the ITU AI for Good awards page, read the new rules when published, and confirm whether the eligibility categories and timeline have changed. Startups and early-stage companies should review the Innovation Factory instead, because ITU specifically directs them there.

For archive readers, the useful lesson from the 2026 call is straightforward: select one category, present a real AI-enabled solution, document measured impact, explain how the work can continue, and be precise about responsible use. The awards rewarded evidence that connected technical work to public value. They were not a substitute for funding, and the closed 2026 deadline should not be presented as an opportunity readers can still enter.

Official reference

The canonical ITU page for this historical cycle is: https://aiforgood.itu.int/ai-for-good-impact-awards/