OLM's New AI Tool Cuts Pokemon Coloring Time by 30%
The studio's system learns visual styles from just three frames, aiming to streamline production without replacing human artists.
By Austin Reed ·

OLM, the Japanese animation company behind the long-running "Pokemon" television series and films, has unveiled a new artificial intelligence tool designed to accelerate the coloring phase of anime production. The system requires only three hand-painted frames to learn a specific production's visual style, a process that takes approximately 90 seconds.
Yotsukura Tatsuo, a board director at OLM Digital, presented the technology at the Taiwan Creative Content Fest in Taipei. He explained that the tool emerged from ANIMINS, a 15-month research project led through GENIAC, a generative AI initiative overseen by Japan's Ministry of Economy, Trade and Industry. The research phase concluded in February.
The primary advantage of the system is its reduced data requirement. Traditional AI models often rely on large training datasets that can be difficult for studios to compile. This new tool instead analyzes the line shapes and color palettes of the three provided frames. It then colors each additional frame in roughly 15 seconds. Yotsukura noted that accuracy improves when artists supply more examples, but the core function remains efficient with minimal input.
Yotsukura clarified that the technology is not strictly generative AI. By using only frames from the specific production, the system avoids interfering with other works. It can also adapt when new characters or creatures are introduced to the series. At the proof-of-concept stage, Yotsukura estimated that the tool has reduced painting workload by approximately 30%. However, he acknowledged that staff responsible for checking every colored frame have faced somewhat more work.
OLM is not the only studio testing these capabilities. Toon Boom Animation, a Canadian software company, is developing similar tools for its Ember AI suite. Harry Ravelomanantsoa, product marketing director at Toon Boom, stated that testing for its coloring tool was nearing completion. An in-betweening tool is expected to enter beta within a couple of weeks. In one trial at a Manila studio, the tool reduced painting time for a scene lasting between 90 seconds and two minutes from 15 hours to two or three hours.
OLM, which employs around 600 people across Japan, Malaysia, Thailand, and the United States, is also experimenting with AI for in-betweening. Yotsukura suggested that larger benefits may arise from supporting productions involving more than 100 people. With a startup partner from the GENIAC project, OLM is exploring AI agents that could help incoming directors and production managers retrieve information about long-running series, such as a director's intentions for a specific cut.
The company is also considering AI agents to speed up approvals and licensing for secondary uses of its titles. Yotsukura described this area as important to revenue but emphasized that OLM is approaching these sensitive applications carefully. He cautioned that professional animation requires reliable results and the ability to make corrections. He noted that even image edits that look acceptable can alter RGB color values, which creates problems for anime production.
Other industry leaders have expressed varying views on the role of AI. Charles Beirnaert, chief creative officer of French studio Blue Spirit, stated that no AI is used on its Netflix series "Blue Eye Samurai." He acknowledged that restrictions might differ among Netflix partners. Netflix's production guidelines, published in August 2025, require disclosure of planned generative AI use and written approval for certain applications.
Beirnaert provided an example where one person used AI to recreate work that had taken about 20 people three months for a conventional CG pilot. While the backgrounds were better, he said the animation was weaker than the original. He maintained that AI should not replace animators. Beirnaert suggested AI could help close financing gaps, estimating €1 million saved on an €8 million budget. Yotsukura countered that OLM joins projects after budgets are set and views AI as a way to expand creative possibilities within those limits.
More Tech · National news · Report a business · Got a Story?






