An emerging US-allied 'AI cold war' dominates the technology outlook as Washington's frontier-model export curbs push Europe and Five Eyes toward sovereign AI, while the Russia-Ukraine front stays in escalation — localized gains near Kupyansk and a near-certain Kremlin disinformation push around the Lavra strike — and Israel holds its post-deal buffer positions in Lebanon and Syria.

The base rate for Russian information operations following high-profile atrocity strikes is near-certain (Bucha, Kramatorsk, Mariupol theatre all followed within days), and the MoD has already publicly committed to the Patriot-blame narrative — the strongest possible seed signal. The Skeptic (risk 83) passed with mild compression to 0.76. The genuine uncertainty is not whether the campaign occurs but whether it is formally DOCUMENTED as coordinated inauthentic behavior within the 30-day window — researcher disclosure timing is variable, though DFRLab/EU DisinfoLab typically publish quickly on flagship events. Critically, my own cybersecurity track record is poor (Brier 0.315, over-predicting by ~26pp), so I deliberately compress toward 50% and weight the documentation-timing risk more heavily, landing at 0.70 rather than the analyst's 0.76. What would make me wrong: researchers deprioritize attribution, platforms suppress detection reporting, or the narrative stays diffuse rather than provably 'coordinated.' Pillars: information_warfare, threat_actor_analysis.