Climate disinformation meets the AI tutor: how foreign actors and Silicon Valley both chase the same classroom

On 14 July 2026, France 24's French-language feed carried a short item that distilled a year of work by the Iris think-tank in Paris: as the climate crisis deepens, foreign actors are increasingly piggy-backing on extreme-weather events to push coordinated disinformation into European information spaces [France 24, 14 July 2026, 15:08 UTC]. The same morning, US financial-newsletter Unusual Whales published a feature on a very different cohort of decision-makers, well-off American parents who are pulling their children out of conventional schools and enrolling them in small, AI-tutored campuses where staff are called "guides" and curricula are tailored by software to each child [Unusual Whales, 14 July 2026, 02:31 UTC]. Read separately, the two stories sit on different shelves: one is a national-security note on information warfare, the other a soft consumer-trend piece on education. Read together, they describe the same contest, fought over the same prize, which is the cognitive environment of the next generation.

The thesis this publication is developing is straightforward. Two distinct pipelines, one state-adjacent and one venture-capital-adjacent, are converging on the classroom. The first floods the public sphere with denial, deflection and false-attribution content timed to disasters. The second quietly reshapes what children are taught, and by whom, inside walls that have little public oversight. Both rely on the same underlying infrastructure: algorithmic curation, large language models and the assumption that the reader, or student, is a passive node to be optimised.

What Iris actually flagged

The Iris report, summarised in France 24's bulletin, does not claim that climate change is a hoax. The framing is more precise and more uncomfortable. As extreme-weather events multiply across Europe and the Mediterranean, hostile state and state-adjacent actors exploit the information vacuum left by the disaster itself. Local communications are down, official sources are overwhelmed and a population seeking answers is unusually receptive to whichever narrative lands first. Into that vacuum, the report says, push actors spreading denial of the underlying science, false attribution of the disaster to a rival power, or simply chaotic content designed to erode trust in any authoritative source [France 24, 14 July 2026, 15:08 UTC].

Iris is the Institut de Relations Internationales et Stratégiques, a French policy institute with a long track record on hybrid threats. The methodological move worth noting is the shift in target: previous generations of climate disinformation aimed at sceptics in the United States and Australia. The current wave, by Iris's account, is aimed at European audiences already broadly convinced of the science. The objective is no longer to convince; it is to demoralise, to make climate policy feel futile, to redirect anger from emitters to elites or to the European institutions carrying the policy.

What the schools story actually describes

The Unusual Whales piece, which originated in a newsletter reaching a finance-and-retail-investor audience, sketches the rise of small private academies serving high-earning US families. Teachers in these schools are re-titled as "guides" or "coaches". AI tutors personalise the curriculum. Class sizes are small. The implicit pitch is that conventional schooling, with its fixed curriculum and unionised teaching corps, is a slow, standardised product unsuited to a labour market in which AI fluency is the binding constraint [Unusual Whales, 14 July 2026, 02:31 UTC].

The educational and ethical questions raised by the model are not new. Personalised learning has been promised by ed-tech vendors for two decades. What is new is the consumer base: families with the capital to opt out of the public system entirely, and a regulatory environment in several US states that allows them to do so with public funds. The story does not name specific schools, founders or funding rounds, and the bulletin flagged it as a soft trend note rather than a verified market map. Treat the institutional detail as directional, not as a roster.

The pattern, stated in plain language

Two things are happening at once. On the public side, foreign-aligned information operations exploit moments of climate stress to push material that corrodes trust in mainstream institutions. On the private side, a parallel, largely commercial operation rebuilds the institution itself, the school, around an AI-mediated pedagogy that determines what each child sees, in what order, and at what pace. Neither actor needs to coordinate. Both benefit from a media environment in which the line between a state-aligned bot network and a venture-backed tutoring platform is harder for an average reader, or parent, to draw.

The structural point is that the contest for cognitive infrastructure is no longer a contest over newspapers, television channels or even social-media feeds in the conventional sense. It is a contest over the algorithms that decide what disaster a teenager sees in the first hour after a flood, and over the algorithms that decide what a seven-year-old is taught on a Tuesday morning. In both cases, the public oversight regime is thin. The European Union's Digital Services Act applies to the first. The second falls into a patchwork of US state-level accreditation rules and private-school exemptions, with the AI tutor itself treated as a product, not a pedagogical actor.

Who wins, who loses, and on what clock

The winners in the first case are the foreign actors named in the Iris framing, principally Russia-aligned and, to a lesser extent, Iran-aligned networks that have a documented history of climate-content experimentation. The losers are European climate policymakers and, more durably, the public's capacity to mount a coherent policy response during the first 72 hours of a climate shock. The winners in the second case are the ed-tech vendors and the small set of wealthy families able to pay the tuition, which the source describes only as "high-earning" without a published figure. The losers are the public-school systems already under fiscal strain, and the broader social contract that depends on a roughly common educational experience.

The time horizon matters. Climate disinformation campaigns are episodic: they spike around a wildfire or a heatwave, then recede. The AI-tutored schools are building a generation-long moat. A child who spends K-12 inside a personalised AI curriculum will, by 18, have a fundamentally different information diet, and a different sense of what counts as authoritative, than a peer in a conventional public school. That divergence is the point. It is also the structural change that will outlast any single disinformation cycle.

What we verified, and what we could not

This publication verified the following from the source material. The Iris report's existence and its general claim about climate-disinformation piggy-backing on extreme weather, as relayed in the France 24 bulletin of 14 July 2026 at 15:08 UTC. The existence of a US trend, reported by Unusual Whales on 14 July 2026 at 02:31 UTC, in which high-earning families are choosing small AI-tutored schools that re-title teachers as "guides" or "coaches" and personalise curriculum via software. The general claim that the two phenomena are occurring in the same news cycle and address overlapping audiences.

This publication could not verify, from the source material, the following and therefore does not assert them. The named authors of the Iris report. The specific size of the AI-tutored school segment in dollar terms. The state-by-state regulatory map. The identity of any specific foreign actor behind the climate-disinformation campaigns beyond the broad Russia-aligned and Iran-aligned categories used by Iris as relayed by France 24. The names of any specific schools in the Unusual Whales piece. Any direct causal link between the two trends; the connection drawn here is structural, based on shared audience and shared infrastructure, and is not asserted as a documented operational relationship.

The open question

The unresolved question, and the one a reader should keep on a watchlist, is regulatory. The EU has begun to treat climate disinformation as a measurable harm under the Digital Services Act, with risk assessments for very large platforms that touch climate content. The United States, at the federal level, has done less. State-level activity, particularly around AI in education, is the active frontier. If 2027 produces a federal floor on AI-tutor disclosure, the small-school segment described by Unusual Whales will be forced to publish what its software teaches, and to whom. If it does not, the divergence between public and private cognitive infrastructure will continue to widen, and the next climate-disinformation bulletin will land on an audience whose default sense of authority is being formed, in real time, by a model whose weights are a trade secret.

Desk note: Monexus framed these two wire items as a single structural story about cognitive infrastructure, rather than running them as separate trend pieces. The Iris report is presented with its sourcing caveat; the Unusual Whales piece is treated as a directional consumer trend, not a verified market map.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/france24_fr
• https://t.me/CryptoBriefing
• https://t.me/france24_fr
• https://t.me/CryptoBriefing