The Astana AI Film Festival (AAIFF 2026) has become the world’s largest international competition dedicated to films created using artificial intelligence, attracting 8067 entries from 125 countries, as reported by the en.kabar.kg.
Kazakhstan ranked second by the number of submissions with 829 films, behind the United States with 1395. India submitted 702 entries, followed by China with 550 and South Korea with 349. The top ten also included the United Kingdom, Indonesia, Japan, Brazil and France.
Nearly 40% of the films, or 3138 entries, explored the festival’s central theme, The Future Worth Living In. Another 4929 works were submitted to the open competition, which covers 19 genres and has no thematic restrictions.
According to AAIFF co-founder and Creative Director Aizatulla Khussain, the idea for the festival emerged in Astana in December 2025 as the organizers anticipated rapid advances in generative video technology.
The entries undergo several stages of screening, beginning with checks for plagiarism and low-quality AI-generated content. Specialists with experience in cinema and visual storytelling assess films that pass the initial review. Around 150 works will reach the festival’s senior selection team, which will choose 25 finalists.
Khussain said hundreds of submissions had produced a strong “wow effect”, making the final selection particularly difficult. He added that two films by Kazakh creators were already among the preliminary frontrunners.
The festival offers a 1 million dollar prize pool, including 750 thousand dollar for its main thematic program and 250 thousand dollar across five open categories. Entrepreneur and investor Balaji Srinivasan to support promising AI teams established a separate 1 million dollar investment fund.
Srinivasan told Qazinform News Agency that falling production costs could create new opportunities for smaller language communities.
The submitted films have a combined runtime exceeding 100 thousand minutes and were created by 4756 filmmakers using more than 80 generative models.
