The AAAI 2026 conference has successfully piloted an AI-powered review system for its Main Technical Track, processing 22,977 papers at a cost of under $1 per review. This initiative, conducted in collaboration with universities and research institutions, marks a significant step in integrating AI into the peer review process. The AI system, leveraging advanced language models, provided detailed reviews that authors and program committee members found more useful than human reviews in several key areas, such as technical accuracy and research suggestions. The AI review system, which does not replace human judgment but supplements it, was able to complete over 20,000 in-depth reviews in less than 24 hours. The system's efficiency and cost-effectiveness were highlighted, with OpenAI providing API resources to support the project. Despite AI's advantages in identifying technical errors and providing improvement guidelines, some concerns remain about its lack of scientific intuition and potential for overanalyzing minor details. Nevertheless, the pilot demonstrates the potential for AI to significantly enhance the peer review process in academia.