Sakana AI Develops Proprietary Technology for SNS Visualization and Countering False and Misleading Information in Ministry of Internal Affairs and Communications Project
Sakana AI ● Covered by 2 sources
Sakana AI built a full misinformation-fighting system for Japan's telecom ministry. It maps, judges, and even test-drives counter-messaging against fake narratives spreading on social media.
Based on reporting by Sakana AI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Tokyo-based Sakana AI just wrapped up a government contract that reads like a blueprint for fighting disinformation at scale. The company was the technology developer chosen for a fiscal year 2025 project run by Japan's Ministry of Internal Affairs and Communications, aimed at building tools that visualize, judge, and respond to false and misleading information circulating online.
The pitch behind the project is straightforward: counting impressions or shares doesn't tell you much when a false narrative is quietly reshaping public opinion. So Sakana AI built what it calls a novelty search technique, one that tracks distinct narratives rather than raw engagement numbers, and pairs it with an automated system that generates analysis reports on how those narratives are spreading through the information space.
Detecting the fakes themselves is the second piece. Generative AI has made convincing fabrications cheap and easy, so Sakana AI stacked multiple detectors together rather than relying on one model's judgment. Crucially, the system also surfaces its reasoning, showing why it flagged something, which matters if actual analysts at ministries or media organizations are expected to trust and verify the calls an algorithm makes.
And then there's the part that goes beyond just spotting problems: a simulation environment for testing counter-messaging before it ever goes live. Instead of guessing whether a rebuttal campaign will calm a spreading rumor or backfire and amplify it, agencies can now run that scenario first. Sakana AI framed all of this as part of a broader push to embed AI into intelligence-related work in Japan, tying it explicitly to national security debates where information capability is increasingly treated as its own strategic pillar.
The results got a public airing on March 16, 2026, at a Ministry of Internal Affairs and Communications event showcasing outcomes from the disinformation countermeasures program, where Sakana AI's team presented to industry players, media, and central government officials. The company says it plans to keep pushing AI adoption in this intelligence space and is recruiting for the effort.
My take — AI-written commentary, not fact-checked reporting
A domestic AI lab building sovereign disinformation-detection infrastructure for its own government is exactly the kind of move I wish more countries outside the US-China axis were making, instead of just importing someone else's black box. That said,
Read more about this at: Sakana AI
Related stories
Sakana AI Wins Award at US-Japan Competition for Defense Innovation
Sakana AI ·
46
Sakana AI Wins Contract for Survey and Proof-of-Concept of AI Functions Needed for Comprehensive Analysis Work from the Ministry of Defense
Sakana AI ·
24
Sakana AI Begins Commissioned Research from Defense Innovation Science and Technology Institute
Sakana AI ·
10