Japan Wants to Take the Profit Out of Election Fakes

Most attempts to fight online election disinformation focus on what people say. Japan’s newest plan targets something more practical: why they say it. Draft guidelines unveiled on Friday by the Ministry of Internal Affairs and Communications aim at the money that flows to accounts which go viral with misleading political content.

What the draft asks of platforms

The guidelines cover 13 major social media operators and spell out how they should apply a revised platform law passed by parliament in July. Among the key points:

• Platforms must set eligibility rules and strict screening for users who want to join monetisation programmes.
• They should suspend monetisation when content that harms the fairness of an election is posted.
• Harmful images and videos generated by artificial intelligence should carry warning labels.

Public comments open on Saturday. The law takes effect on 1 March 2027, in time for Japan’s unified local elections next spring.

Following the incentive

The focus on monetisation reflects a pattern seen in recent Japanese campaigns, where some accounts discovered that outrage, rumours and clipped videos of candidates could earn real advertising income. In that model, a false claim is not only a political weapon but also a business. Cutting off revenue is a way to change behaviour without the government deciding what counts as true, a line Japanese officials are careful not to cross because of strong constitutional protections for expression.

The approach also sidesteps a hard technical problem. Detecting every bot or coordinated network is nearly impossible, but checking who is allowed to earn money is something platforms already do for tax and fraud reasons. Tightening that gate is cheaper and more measurable.

Open questions

Critics will ask who decides when a post “harms fairness”, and how quickly platforms will act during the few intense weeks of a campaign. Smaller local candidates may worry that rules designed for national debates could be applied unevenly. And AI labels only work if people notice them and if the tools that create fakes do not outrun the tools that detect them.

Still, Japan’s draft offers a model others in Asia are watching. South Korea, Taiwan and Indonesia have all wrestled with deepfakes and paid influence operations around their votes. If Tokyo can show that removing the profit motive reduces the noise without silencing legitimate voices, the idea may travel far beyond its local elections.