A warning that Britain's information defences aren't keeping up, a World Cup swarmed by deepfakes, and the leaked files inside Russia's influence machine in Africa

Britain's fact checkers say the country's defences aren't keeping up

Full Fact published its seventh annual report this week. The charity argues the UK's information environment should now be treated as critical democratic infrastructure that no single institution is responsible for defending, and that the safeguards in place are not keeping pace with how people actually find things out.

The throughline is AI. More than half of UK adults now use tools like ChatGPT or Gemini, and Full Fact's polling found only 3% feel very confident telling a genuine video from a synthetic one. Search overviews and chatbots increasingly sit between people and original sources, deciding which claims get foregrounded and which get left out. Since January the charity has run its own daily benchmark, feeding leading models the same questions to see how consistent and well-sourced their answers stay over time.

The report arrives with a warning from the Electoral Commission, whose chief John Pullinger contributed a guest essay arguing platforms must own the abuse spreading in their spaces. Full Fact's recommendations run to institutional reform: a national information incident framework, systemic-risk duties for platforms and AI systems, and provenance labelling. The Cabinet Office refreshed its crisis doctrine last year. The report notes it still says little about what to do when the contested thing is the information itself.

A World Cup swarmed by deepfakes

The 2026 World Cup has drawn a wave of AI-generated fakes running from harmless novelty to outright fraud. On 11 June, the day of the opening match, a YouTube channel livestreamed an entirely AI-generated "2026 World Cup Opening Ceremony" that pulled in more than 1.42 million cumulative views, complete with tells like an Argentine flag and an elephant on the pitch. A real ceremony took place the same day in Mexico City; this was a separate, fabricated broadcast.

The week's most-shared fake deceived millions: a clip of Erling Haaland startled by his own reflection in a restaurant drew more than 31 million views on X within days. The footage was an AI face swap built on a comedy skit by Chinese comedian Jin Long, Wired reported, and it kept spreading after fact-checkers traced the original.

The larger problem is commercial. The deepfake-detection firm Revelum says it has tracked more than 10,000 scam adverts impersonating footballers over the past year, promoting fake gambling apps, investment platforms and crypto schemes. Activity climbed once the tournament began: the firm recorded a roughly 1,700% rise in Neymar-branded scam ads after 11 June. One ad even claimed a Neymar deepfake had been created by competitors to discredit a legitimate platform, turning public awareness of deepfakes into a prop for false trust.

Around the tournament's edges, FortiGuard Labs counted more than 13,000 FIFA-themed domains registered between January and May, with about 8.8% flagged as malicious or suspicious. The tournament runs to 19 July.

A network of fake brides is waiting for Taiwan's November vote

NewsGuard mapped a cluster of 294 coordinated accounts on Meta's Threads, launched in May, that pose as attractive women looking to marry Taiwanese men. The accounts share posting schedules, handle patterns, and a fixation on Taiwanese men. About 40% follow the naming convention of a network that pushed attacks on the pro-Taiwan Democratic Progressive Party earlier this year.

Seventy of the accounts posted identical dating profiles. Several get Taiwanese hobbies and food wrong: one described Hsinchu rice noodles as sweet and sour, though the dish is savoury. As of late June none had posted political content. NewsGuard assessed the network looks poised to switch ahead of Taiwan's local elections in November, when voters choose leaders whose stance toward Beijing is at issue.

Using appealing women as a lure is an old move in Chinese operations. AEI's Taiwan trackers reached the same read from a different angle in late June, assessing that Beijing is likely running fake social accounts to sway the vote by amplifying local voices aligned with its aims. The network is already built and, so far, has stayed quiet on politics.

The leaked files behind Russia's covert machine in Africa keep widening

A consortium led by The Continent and Forbidden Stories, with Dossier Center, iStories, openDemocracy and All Eyes on Wagner, has been working through 1,431 pages of internal documents from a network called "the Company," better known as Africa Politology. European security sources authenticated the files. They describe an influence architecture run by Russia's foreign intelligence service, the SVR, which absorbed the operation from Yevgeny Prigozhin's Wagner orbit after his death in 2023.

Before Namibia's 2024 election, operators pushed a forged letter claiming the UK was secretly financing the opposition. In South Africa, spreadsheets show local influencers paid roughly $150 a post to attack the Democratic Alliance under campaign names like "DA Racists." In the Central African Republic, outlets such as Lengo Songo laundered the messaging, and a separate chapter documented by openDemocracy budgeted around $283,000 to place articles across more than twenty Argentine sites. The files also sketch coup-scenario planning in Senegal and a "Confederation of Independence" project to build a belt of Moscow-friendly regimes from the Atlantic to the Red Sea.

The spreadsheets seldom show whether the money was actually paid, or to whom. Some entries repeat. And because operators are rewarded for results, the reporters caution, the figures may be padded to win the next contract. The budgets, at least, are set out in full. In 2024 the network's African operations were planned at more than $7 million, with the Central African Republic the largest single line.

Bombs and brochures: Russia's dual-use drones over Kherson

Leaflets dropped from drones are not a new feature of the war; Russia has been pairing its FPV drones with propaganda for some time. What an analysis published by EUvsDisinfo on 29 June adds is a closer look at how the tactic works. In Kherson, drones flying from the east bank of the Dnipro drop leaflets over the city, the same type of drone used to drop explosives on its residents.

The analysis describes a sequence: strikes and attacks on infrastructure wear residents down, telecommunications are damaged so propaganda channels fill the vacuum, and the leaflets then arrive with offers of safety for those who accept the occupier's version of reality. In Nikopol, in the Dnipropetrovsk region, drones struck a television tower and cut the broadcast signal before a leaflet campaign began, with Russian radio filling the gap.