Guest Essay – Digital Electoral Interference in Japan: Lessons from Okinawa
By Masakazu Takamori and Jonathan Berkshire Miller
When Genta Koja, the 42-year-old former vice mayor of Naha, defeated two-term incumbent Denny Tamaki in Okinawa’s September 13 gubernatorial election, the result was immediately read through a single lens: a referendum on the U.S. military presence that has defined Okinawan politics since reversion in 1972. Within hours, a second narrative appeared alongside it. Commentators, on both sides of the outcome, began asking whether social media had shaped, or distorted, the result.
This is not a uniquely Japanese impulse. It is now the standard post-election reflex in democracies everywhere, Canada included. And it deserves more scrutiny than it usually gets.
Japan’s information space has been shifting for years, but the pace of that shift keeps outrunning expectations. Warnings issued at NATO’s Strategic Communications Centre of Excellence conference in Riga in 2024, after case after case of Russian interference in European elections, held that similar dynamics would reach Japan within a year or two. In practice, it took about six months. The direction was right. The pace was not: it moved faster than most analysts anticipated.
That speed is the real story of Okinawa 2026, more than any particular claim about who benefited from what post.
What actually happened in Okinawa?
Koja won convincingly, taking 59 percent of the vote to Tamaki’s 38 percent, a margin of more than 147,000 votes and the largest tally for any Okinawa gubernatorial candidate since 1972. Turnout rose to roughly 61 percent, about four points higher than 2022. These are not the numbers of a result manufactured by a viral post or a coordinated bot network. They reflect a genuine and substantial shift in Okinawan preferences.
Pre-election polling supports that reading. A Ryukyu Shimpo and JX Press survey conducted in late August found that nearly 47 percent of respondents cited the economy, jobs and rising prices as their top concern, against 16.6 percent who named U.S. bases and security. Analysts on the ground described the two candidates’ economic platforms as barely distinguishable, which meant the campaign turned on questions of generational change and administrative competence as much as on the long-running Futenma dispute.
That did not stop the information space from being active. In late July, a false post claiming Tamaki had already agreed to relocate the Marine Corps’ Futenma air station to Henoko, and that his opposition campaign was a publicity stunt, spread widely on X. A Jiji Press analysis published after the vote found that almost all reposts of that claim came from accounts outside Okinawa. Tamaki himself threatened legal action over what he called defamatory online harassment during the campaign.
So, the ingredients for a “social media decided the election” narrative are all present: viral falsehoods, out-of-prefecture amplification, an emotionally charged security issue. What is missing is evidence that any of it moved the result by a margin that mattered, when a 21-point economy-driven landslide was already underway.
The caution both sides should observe
None of this means social media is irrelevant to Okinawan or Japanese elections. It plainly is not, and its influence is intensifying faster than institutions can build the capacity to fact-check it in real time within a short campaign period. The point is narrower: the label of a result “distorted by social media” should not be applied before the evidence supports it, and it is a label available to winners and losers alike as a convenient explanation for outcomes they would rather not accept on their own terms.
The sharper risk sits one level below the question of who won. Influence operations, whether foreign-directed or domestic, rarely need to flip a vote count to succeed. Their more durable objective is to deepen social division by making every contested issue, and eventually every election result, appear illegitimate to a meaningful share of the population. That erosion of confidence in the process itself is harder to measure than a vote count and harder to reverse. It deserves attention before, not after, the horse race of who “really” won.
Part of the difficulty is evidentiary. Obtaining conclusive evidence of coordination, attribution, or actual impact on social media is extremely difficult, which is exactly why the data and the surrounding context both call for careful examination, with a clear line drawn between what the evidence demonstrates and what can only be offered as a possibility.
Why this matters for Canada
Canada has its own recent experience with the difficulty of separating genuine foreign interference from overstated claims about it, through the Hogue Commission’s work on the 2019 and 2021 federal elections. Commissioner Marie-Josée Hogue’s final report found only a very small number of cases where foreign meddling may have affected nomination races or individual ridings, with no evidence that it altered which party formed government. That inquiry illustrated the same lesson Okinawa now offers: attribution takes time, evidence discipline matters more than speed, and premature conclusions, in either direction, damage public trust regardless of which way they lean.
There is also a strategic dimension. Okinawa’s political direction affects the pace of Futenma’s relocation to Henoko and the broader U.S. force posture in the East China Sea, developments Ottawa tracks closely as Canada deepens defence industrial cooperation with Japan and expands its presence under the Indo-Pacific Strategy. A Japan more settled on its southern security architecture is a more predictable partner for Canada on Indo-Pacific security cooperation, from naval deployments to critical infrastructure protection. Understanding what drove the Okinawa result, rather than defaulting to the social media explanation, is directly relevant to how Canadian policymakers read Tokyo’s trajectory under Prime Minister Sanae Takaichi’s government.
The information space around elections is only going to become more contested, in Naha, in Ottawa, and everywhere in between. The right response is not to dismiss its influence but to insist on the same standard of evidence applied to any other causal claim about a democratic outcome: show the mechanism, show the scale, and show that it changed minds rather than confirmed them. Until that work is done, restraint is not a lack of interest in the problem. It is the discipline the problem requires.
Masakazu Takamori is the President and CEO of Japan Nexus Intelligence. Jonathan Berkshire Miller is principal at Pendulum Geopolitical Advisory and Senior Fellow at the Montreal Institute for Global Security.