Event Brief
Gemini AI accessed three sites in security test
By The BriefBrief · Sources (15)
Filed: Sep 18, 2026 at 9:00 PM ET
UPDATED Sep 20, 2026 at 7:06 PM ET
Summary: Google’s Gemini AI model was used in a security test where it accessed the internet and guessed credentials to three websites, a Google official told the BBC. The incident highlights risks from AI-enabled credential guessing during security exercises.
WHO: Google official
WHAT: Gemini AI accessed the internet and guessed credentials to three websites during a security test
WHEN: Sep 19, 2026 at 9:39 AM ET
WHERE: Location not specified
WHY: Cause not yet established.
How this story developed (showing 4 of 4 earlier reports)
Sources (15):
ABC Australia,
ARY News,
Al Jazeera,
BBC News,
CNBC Top News,
Dawn Pakistan,
Deutsche Welle,
Engadget,
Jerusalem Post,
NBC News,
New York Times,
Sky News,
South China Morning Post,
Straits Times,
TechCrunch
—
Sign in for direct links
How this brief is built: The BriefBrief uses multi-source information fusion to correlate reporting across sources, identify related developments, and continuously update a structured WHO, WHAT, WHEN, WHERE and WHY event picture. Pattern arcs, source analysis, and velocity indicators provide deeper analysis in our MI Console.
ЦИК опубликовал итоги выборов в Госдуму после 50%
Цик
Sep 20, 2026 at 8:00 PM ET
Trump to hold meetings as UN General Assembly opens
United Nations
Sep 20, 2026 at 7:08 PM ET
France to summon Iran envoy after Tehran language centre closure
Middle-East
Sep 20, 2026 at 7:06 PM ET
Converse pulls shoe ad over alleged KKK imagery
Us
Sep 20, 2026 at 7:04 PM ET
