Japan will require AI-generated images and videos circulated on social media during election campaigns to carry clear labels under legislation that cleared parliament this week in response to a string of controversies over synthetic content used to attack political rivals.
Internal Affairs Minister Yoshimasa Hayashi confirmed the new rules on Tuesday, telling reporters that safeguarding the integrity of Japan’s elections made the measure essential.
The rules were approved by parliament on Monday and are due to take effect in March and are aimed at protecting the integrity of the country’s electoral process.
Beyond the labeling mandate, the legislation also bars both individual internet users and platform operators from circulating fabricated or manipulated claims about candidates.
Notably, the disclosure requirement is not blanket. According to reporting from the Japan Times, the labeling mandate is aimed at synthetic videos and images realistic enough to be mistaken for authentic footage, while content that is obviously AI-made will be exempt from the rule.
The distinction is meant to target deceptive deepfakes specifically, rather than sweeping in every meme or AI-assisted graphic that campaigns and supporters routinely post.
The push for regulation traces directly back to Japan’s political turmoil of the past year. The rules follow allegations about the use of AI-generated content to smear candidates in a 2025 leadership contest in the ruling Liberal Democratic Party and in parliamentary elections in February, a reference to the LDP presidential race that ultimately installed Sanae Takaichi as the country’s first female prime minister.
That contest was rocked by revelations from an IT entrepreneur, Ken Matsui, who told Kyodo News he had produced roughly a thousand to fifteen hundred short AI-generated videos attacking Takaichi’s rivals, including Shinjiro Koizumi and Hayashi himself—the same Hayashi who now, as internal affairs minister, is charged with enforcing the new anti-disinformation rules.
Matsui said he was approached by an aide to Takaichi seeking ways to blunt Koizumi’s early lead in the race and that the videos were then pushed out through several hundred anonymous social media accounts. Takaichi has denied that her campaign was involved in producing or distributing the material.
Takaichi went on to win the LDP leadership in October and led the party to a strong showing in February’s lower house election, a victory partly credited to her campaign’s social media reach, adding an uncomfortable backdrop to the debate over how AI content shaped the vote.
Perhaps the most consequential detail is what the legislation does not do. Unlike the European Union’s Digital Services Act, which empowers regulators to levy substantial fines on platforms that fail to police harmful content, Japan’s new framework carries no penalties for non-compliance.
That omission has already drawn skepticism from domestic commentators, with some local outlets openly questioning whether a rule with no enforcement mechanism can meaningfully change platform behavior.
The Nikkei business daily has floated one explanation: that Tokyo may be trying to sidestep the kind of transatlantic friction that has flared between Washington and Brussels over EU fines against American tech companies under the Digital Services Act. By opting for a softer, guidance-based approach, Japan may be attempting to tighten oversight of AI content without inviting similar diplomatic or trade tensions with the United States, home to most major platform operators.
In place of fines, the government plans to issue guidelines instructing platform operators on how to comply, paired with a system of annual disclosures on how those guidelines are being implemented, according to Kyodo News.
Whether that lighter-touch model, essentially relying on reputational pressure and routine reporting rather than legal sanction, proves sufficient to deter bad actors remains an open question that will likely shape debate ahead of the law’s March implementation.
Government officials involved in drafting the rules have framed the effort as a balancing act. As officials have said, in drafting the new rules, they needed to strike a balance between free speech and the sanctity of democratic elections language that reflects Japan’s broader caution around content regulation, given the country’s strong constitutional protections for expression and a political culture wary of heavy-handed state intervention in media.
With national and local elections on the horizon, and with the Matsui affair still reverberating through Japanese politics, the coming months will test whether labeling requirements and voluntary platform guidelines are enough to contain the kind of synthetic disinformation that arguably helped shape the LDP’s last two electoral contests or whether Tokyo will eventually be forced to follow Brussels’ lead and put real penalties behind the rules.
WHAT YOU SHOULD KNOW
Japan has passed new election rules requiring realistic AI-generated images and videos to be labeled and banning the spread of fake information about candidates, prompted directly by revelations that AI-made attack videos were used against rivals in the LDP leadership race that brought Sanae Takaichi to power.
But the law has no penalties. Unlike the EU’s Digital Services Act, Japan is relying on voluntary guidelines and annual disclosures rather than fines, meaning its real-world impact will depend entirely on whether platforms choose to comply, not on any legal obligation to do so.
