Edition 17 · October 16, 2026
UK Prime Minister Andy Burnham said it is “clear” Russia interfered in the 2016 Brexit vote, the first time a sitting prime minister has said so. MI5’s chief backed the claim but said there is no proof it changed the result. Bloomberg reports Russia, Iran, and China are all using AI to push disinformation before the November 3 US midterms. A viral story that Ukrainian strikes on Russian data centers had knocked out Kremlin bot networks turned out to be wrong. It rested on misread traffic data. France’s foreign minister, meanwhile, accused Elon Musk’s X of running harassment campaigns against European leaders.
Brexit, Ten Years On
- PM Andy Burnham said it is “clear” Russia interfered with the Brexit vote (BBC, Oct 11), warning of similar activity in France and Germany; it is the first such confirmation by a sitting UK prime minister.
- MI5 chief Ken McCallum confirmed Russian interference attempts via media and fake accounts (News18, Oct 13), but said it does not prove Russia changed the outcome.
Election Integrity
- Bloomberg reports Russia, Iran, and China are using AI to boost disinformation ahead of the US midterms (Bloomberg, Oct 14) on November 3.
- New detail on the Russian operation OpenAI flagged last week: it used spoofed emails, fake government documents, and front organizations to target schools in Peru and Ecuador (Dark Reading, Oct 14).
- Iran’s seven AI-generated fake journalists pitched articles to outlets worldwide (Jerusalem Post, Oct 12), per OpenAI, not only in the US.
The Bot Story That Wasn’t
- President Zelensky told the Warsaw Security Forum that strikes on Yandex’s data centers made social media “a little easier to breathe.” Within hours the claim had grown specific: MAGA, Reform UK, and AfD accounts had supposedly gone quiet, and one post claiming AfD bot activity fell 56% was viewed millions of times before being deleted (Mother Jones, Oct 13).
- What actually dropped was Yandex’s web crawlers. The widely shared Cloudflare chart showed indexing traffic falling, not troll accounts; as Georgetown’s Renée DiResta notes, “bot” means both, and trolls “usually have more resilient infrastructure.”
- The numbers did not move. Common Consultancy compared 76 AfD politicians and 136 MAGA accounts on Facebook against their own baselines and found ordinary noise (AfD −8% to +5%, MAGA +4% to +10%), while Antibot4Navalny reported the Kremlin-linked Matryoshka network still posting the next day.
- The BBC rated the claim unsubstantiated (BBC, Oct 13), and Süddeutsche Zeitung traced the viral AfD figure to a false AI-generated graphic (SZ, Oct 13).
- Mother Jones argues people were primed to believe it: a 2023 Indiana University study found users overestimate how many bots there are, and see their political opponents as the ones being fooled by them.
- What we saw: the Cloudflare chart proves neither side. Aggregated network data shows traffic volume by hosting provider, not who is behind the accounts. What we observed is a few tracked bots going dark for a brief period. A real influence network is spread across accounts, hosting, and people, so a strike on one provider’s data center is not the kind of event you would expect to show up as a clean drop in bot activity, especially for long.
Platforms Under Pressure
- French FM Jean-Noël Barrot accused Elon Musk’s X of “online harassment campaigns” against European leaders (The Guardian, Oct 14) ahead of the 2027 election.
- Microsoft’s X account, with 13M+ followers, was hacked to promote a fake “$Clippy” cryptocurrency (Fox News, Oct 11); Ireland’s Garda police accounts were hijacked for a crypto wallet scam (TheJournal, Oct 15).
- Turkey’s Halk TV went dark after a police raid (SBS, Oct 14), as rights groups urged platforms to resist Turkish orders blocking 1,270+ journalist and activist accounts (US News, Oct 12).
Espionage & Recruitment
- Australian police allege Chinese intelligence targeted 100+ Australians, including a former premier, via LinkedIn (ABC, Oct 14), posing as think tanks offering paid “advice”; Beijing called it a smear.
Deepfakes
- Nigeria’s defence chief warned AI fake videos and cloned voices could endanger troops and trigger panic (Guardian Nigeria, Oct 13).
- President Trump posted a fake quote attributed to Denzel Washington praising him (WCVB, Oct 12), a line previously misattributed to other celebrities.
- CU Boulder researchers developed a lighter method to detect deepfakes in compressed video (CU Boulder, Oct 12), aimed at video-call scams.
- The European Environment Agency’s director warned AI climate disinformation is out-distributing verified data (Tech Times, Oct 14).
India
- Indian intelligence alleges a Pakistan-origin network, “Bharat Bytes,” used fake media identities to spread false police-violence claims (Moneycontrol, Oct 13) during Delhi protests; the attribution is a government claim.
- Fact-checkers confirmed viral videos of women protesters were AI-generated (The Quint, Oct 13), verified with SynthID and Hive.
Quick Hits
- In Ethiopia, social media has become a propaganda battleground (France 24, Oct 13) as conflict resumes in the north and journalists lack access.
- Euronews asks whether a new European Security Council can counter Russia’s hybrid attacks (Euronews, Oct 15); no legislative proposal has been tabled yet.
- A viral claim that Trump apologised to Nigeria’s president was debunked (Legit, Oct 13); the screenshot was AI-altered.
About Bad Information
A weekly overview on disinformation, coordinated influence operations, and media literacy. Researched and fact-checked with Aunoo.
Powered by Aunoo — news intelligence with trust built in. Aunoo scores every article on five trust signals before you even click: source bias, factual-reporting rating, AI-content detection, disinformation-network flags, and a quality score. You see how far to trust a story, not just what it claims.
Aunoo is the context layer for narrative intelligence. Try it, it’s free → agentic.aunoo.ai
AI Transparency
This newsletter is produced with AI. Aunoo uses automated systems to gather, score, and draft the content, with human review before publication. In line with the EU AI Act’s transparency obligations (Article 50), which took effect August 2, 2026, we disclose that parts of this edition were generated or assisted by artificial intelligence. Every claim links to its original source — please verify against those before reuse.
