Thursday, October 22, 2026
Deepfakes (4)
- Brazil's electoral court orders Meta to account for a pro-Lula WhatsApp channel as Lula camp alleges AI bias
Courthouse News
Meta says its image tool refused Lula images and was fixed Oct 16; court set a 500,000-reais daily fine for unequal treatment. - PolitiFact ties an AI-generated video smearing Texas Senate candidate James Talarico to Russia's Storm-1516
PolitiFact · 17:45 UTC
PolitiFact ties an AI-generated video making a baseless accusation against Texas Senate candidate James Talarico to Russia's Storm-1516 operation. - Spain's Partido Popular posts an AI video of the PM's wife and allies 'confessing'; PSOE threatens to sue
The Olive Press · 16:45 UTC
51-second clip with AI likenesses of Begoña Gómez, Zapatero and Ábalos; disclaimer in tiny lettering, per El País. - Incheon court orders a former student to pay 17.7 million won over deepfake sexual images of a teacher
Seoul Economic Daily · 03:00 UTC
Incheon District Court orders a former student to pay 17.7 million won to a teacher over deepfake sexual images.
The flaw under most websites (11)
- Exploit attempts in the millions as South Korea raises its cyber alert and governments rush to patch
Reuters, Korea Herald · 23:06 UTC
Reuters reports exploit attempts in the millions as South Korea raises its private-sector cyber alert and activates emergency response steps. - NCSC warns more than 50,000 UK organisations; 32 government services offline as a precaution
ITV News · 22:27 UTC
NCSC alerted more than 50,000 UK organisations; a minister told the Lords 32 government services were offline as a precaution. - India's CERT-In rates the PHP flaws critical; no Indian victims identified so far
The Free Press Journal · 15:48 UTC
CERT-In rated the PHP flaws critical in advisory CIVN-2026-0510; no Indian victims identified so far. - Australian tax, health and ASD sites went offline over the PHP flaw; minister says 'this isn't a breach'
The Canberra Times · 15:46 UTC
ASD, ATO and health department sites went offline over the PHP flaw; Science Minister Tim Ayres says 'this isn't a breach'. - Fortune: the PHP flaws may have been exploited before disclosure; OpenAI expands its Daybreak cyber program
Fortune · 14:37 UTC
Fortune reports hackers may have exploited the PHP flaws before disclosure and that OpenAI is expanding its Daybreak cyber-defense program. - Germany, the Netherlands and Japan confirm no successful PHP2Shell exploitation despite millions of attempts
The Next Web · 13:03 UTC
Officials in Germany, the Netherlands and Japan confirm no successful PHP2Shell exploitation despite millions of attempts reported by Cloudflare's CEO. - KELA: cybercrime forums buzz over PHP2Shell, with unverified exploit sales and doubts a reliable one exists
KELA · 12:47 UTC
KELA finds cybercrime forums buzzing over PHP2Shell, with unverified exploit sales and skepticism that a reliable exploit is circulating. - Canada's CIO: 'no indication' federal systems were compromised; Statistics Canada's site back online
Global News · 10:26 UTC
Canada's CIO says there is no indication federal systems were compromised; Statistics Canada's site returned around noon Thursday. - Il Sole 24 Ore: how the early-leaked PHP flaws pushed European institutions to pull websites offline
Il Sole 24 Ore · 10:14 UTC
Italy's business daily describes how the early-leaked PHP2Shell flaws pushed European institutions to pull websites offline. - AWS saw PHP2Shell probing four days before the patches; over a million servers still unpatched on Oct 21
SecurityWeek · 09:48 UTC
Amazon honeypots saw PHP2Shell probing on Oct. 17, four days before patches; over 1 million servers remained unpatched Oct. 21. - FBI and NSA urge organizations to patch the PHP flaws, take unpatched systems offline and hunt for compromise
Newsweek · 04:42 UTC
FBI and NSA urge organizations to patch the PHP flaws, take unpatched systems offline and check for signs of compromise.
Jobs (2)
- Goldman economist sees up to 9% of workers displaced by AI, about one in 11 counting slower hiring
Fortune · 18:20 UTC
Goldman's Joseph Briggs estimates AI displaces up to 9% of workers, plus 3% from slower hiring, with losses concentrated in recessions. - Barclays to cut about 5,000 jobs, 5% of staff, with 60% of cuts in support roles including developers
Reuters
Reuters: Barclays will cut roughly 5,000 roles this year, about 5% of staff; CFO says 60% fall in support functions including developers.
AI attacks (5)
- Wiz: an ARTEX-driven actor overlapping Storm-0940 password-sprays software engineers in 24-to-30-hour sessions
Wiz Research
Hundreds of Entra ID and Okta users hit; ARTEX had 12,000-plus forks when it went closed-source Oct 8; no victim named. - CrowdStrike: four hacking crews ran Claude Code agents largely on their own; one hit 15 Argentine clinics
Cybersecurity Dive · 14:35 UTC
Agentic intrusions roughly tripled in the second half of 2026; a Russian-speaking actor gave its agent 'near-total autonomy'. - Microsoft pins the ConnectWise attacks on China-aligned Storm-0940 using ARTEX AI agents against 2,250 tenants
The Record
Microsoft attributes the ConnectWise supply-chain attack to China-aligned Storm-0940, which used the ARTEX AI agent framework against 2,250 tenants. - Malicious LiteLLM release hit PyPI and Docker Hub with a credential stealer; leaked GitHub tokens suspected
The Hacker News
A malicious LiteLLM 1.107.0 release hit PyPI and Docker Hub with a credential stealer; LiteLLM suspects leaked GitHub tokens in gateway logs. - 404 Media: AI-found bugs like PHP2Shell show how fragile software is, but most AI-written exploits still fail
404 Media · 10:23 UTC
Jason Koebler argues AI-found bugs like PHP2Shell are exposing how fragile software is, while most AI-generated exploits still fail.
AI incidents (3)
- AI aggregator 302.ai left 17.6 million files and 144,000-plus chat logs in an open Azure bucket for a month
TechRadar · 15:35 UTC
Cybernews found 22.72TB of prompts, replies, names, birth dates and corporate documents; the Hong Kong owner never replied. - Outside evaluators found holes in OpenAI's agent monitoring, one letting agents hide hundreds of actions
Reuters, KSL · 10:02 UTC
METR and Redwood Research found weaknesses in OpenAI's internal monitoring, including one that could let agents hide hundreds of actions. - Four big US banks found unvetted third-party AI agents inside customer accounts and are restricting them
Axios · 09:00 UTC
Morgan Stanley, Bank of America, JPMorgan and Wells Fargo found unvetted third-party AI agents in customer accounts and are restricting them.
Families and kids (1)
- A $55,000-a-year AI school says students master core subjects in two hours a day with no teachers
CBS News · 12:30 UTC
CBS visits Alpha School's Chicago campus; a University of Colorado analysis found about a third of its adult guides have education training.
Breakthroughs (3)
- Anthropic says Claude wrote formal proofs for 123 open-source libraries and found hundreds of defects
Anthropic
Anthropic says Claude wrote formal proofs for 123 open-source libraries and found hundreds of defects, some undetected for years. - Microsoft's diagnostic AI scored 83% on 304 hard NEJM cases in Science, against 36% for generalist doctors
Microsoft AI
Science paper reports MAI-DxO reached 83% accuracy on 304 NEJM cases versus 36% for physicians with references; a Beth Israel trial follows. - Google DeepMind's ForTy map sorts every 10-metre patch of Earth into six forest types with 90% accuracy
Google DeepMind
Built on AlphaEarth embeddings with WRI and IIASA, validated at 9,000-plus sites, published in Scientific Data and free on Earth Engine.
Other (1)
- WSJ: OpenAI fired three safety researchers after checking their phones for Signal chats with an outsider
The Wall Street Journal
The three, who had pushed to keep model reasoning visible to outsiders, deny breaking rules; OpenAI reportedly cited a breach of trust.
Headlines are written by Warning Shots and link straight to the outlet. Later corrections and retractions are recorded on the incident pages and in the changelog.