Google’s Gemini hacked three companies in new AI safety incident - Financial Times

Google’s Gemini hacked three companies in new AI safety incident

* Sign In
* Subscribe

Open side navigation menuOpen search bar

SubscribeSign In

* Home
* World
Sections

+ World Home
+ Middle East war
+ Global Economy
+ UK
+ US
+ China
+ Africa
+ Asia Pacific
+ Emerging Markets
+ Europe
+ War in Ukraine
+ Americas
+ Middle East & North Africa
Most Read

+ Trump says US has deal with Denmark for ‘control’ of Greenland’s security
+ UK’s top taxpayer says he would ‘not wish to be reborn’ in Britain
+ The Anglo-French moment
+ Unpacking the real fiscal costs of immigration
+ Investors warn Anthropic could struggle to sustain revenues post-IPO
* US
Sections

+ US Home
+ US Economy
+ US Companies
+ US Politics & Policy
Most Read

+ Trump says US has deal with Denmark for ‘control’ of Greenland’s security
+ Investors warn Anthropic could struggle to sustain revenues post-IPO
+ Trump bans major US news outlets from White House
+ Warren Buffett’s son to succeed his father as Berkshire Hathaway chair
+ Trump says son paid back Russian businessman for wedding weekend gift
* Companies
Sections

+ Companies Home
+ Energy
+ Financials
+ Health
+ Industrials
+ Media
+ Professional Services
+ Retail & Consumer
+ Tech Sector
+ Telecoms
+ Transport
Most Read

+ UK’s top taxpayer says he would ‘not wish to be reborn’ in Britain
+ Google’s Gemini hacked three companies in new AI safety incident
+ Investors warn Anthropic could struggle to sustain revenues post-IPO
+ OpenAI expects to burn $280bn by 2030
+ Trump bans major US news outlets from White House
* Tech
Sections

+ Tech Home
+ Artificial intelligence
+ Semiconductors
+ Cyber Security
+ Social Media
Most Read

+ Google’s Gemini hacked three companies in new AI safety incident
+ Investors warn Anthropic could struggle to sustain revenues post-IPO
+ OpenAI expects to burn $280bn by 2030
+ Oracle’s $18bn data centre debt under strain amid local pushback
+ OpenAI breached by researchers using Anthropic models
* Markets
Sections

+ Markets Home
+ Alphaville
+ Markets Data
+ Private markets
+ Equities
+ Bonds
+ Currencies
+ Commodities
+ Crypto
+ Monetary Policy Radar
+ Wealth Management
+ Moral Money
+ ETF Hub
+ Asset management
Most Read

+ The Iran energy shock is wrongfooting the world
+ PizzaExpress puts £500mn sale on the menu
+ How should investors position for the robot apocalypse?
+ Yen sinks after BoJ raises rates to highest level since 1995
+ Ex-Apollo executive admits to improperly using company documents
* Climate
* Opinion
Sections

+ Opinion Home
+ Columnists
+ The FT View
+ The Big Read
+ Lex
+ Obituaries
+ Letters
Most Read

+ The Anglo-French moment
+ Unpacking the real fiscal costs of immigration
+ The relentless creep of Britain’s compliance state
+ How should investors position for the robot apocalypse?
+ Anthropic and the golden rules of business
* Lex
* Work & Careers
Sections

+ Work & Careers Home
+ Business School Rankings
+ Business Education
+ Europe's Start-Up Hubs
+ Entrepreneurship
+ Recruitment
+ Business Books
+ Business Travel
+ Working It
Most Read

+ Startup Nscale’s CEO pay nearly matched the FTSE 100 record, because of course it did
+ Thousands of Barclays staff revolt against three-day office mandate
+ Who was right on job creation — Adam Smith or Schumpeter?
+ Exclusive: a first look at The Horses, London’s most keenly awaited new pub
+ How to make good decisions under pressure
* Life & Arts
Sections

+ Life & Arts Home
+ Arts
+ Books
+ Food & Drink
+ FT Magazine
+ House & Home
+ Style
+ Puzzles
+ Travel
+ FT Globetrotter
Most Read

+ The Anglo-French moment
+ I’ve covered wine for 50 years. Here are 50 ways it changed
+ Blockbuster Ed Sheeran tour in doubt after fallout from protests
+ Jemima Kelly: my week at Rafa Nadal’s school of padel
+ Raspberry Pi founder Eben Upton: ‘I’m an omni-geek’
* How To Spend It
* Stocks Game

MenuSearch

* Home
* World
* US
* Companies
* Tech
* Markets
* Climate
* Opinion
* Lex
* Work & Careers
* Life & Arts
* How To Spend It
* Stocks Game

Financial Times

SubscribeSign In

Google LLC

Manage your delivery channels here

Google’s Gemini hacked three companies in new AI safety incident

Breakout during training exercises follows instances at rivals OpenAI and Anthropic as concerns grow over safety of frontier systems

!Demis Hassabis stands on stage presenting in front of a large screen displaying "Gemini Omni Flash" and product availability details.

*Demis Hassabis, chief executive officer of Google DeepMind

© Bloomberg*

* Google’s Gemini hacked three companies in new AI safety incident on x (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on facebook (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on linkedin (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on whatsapp (opens in a new window)
*

* Google’s Gemini hacked three companies in new AI safety incident on x (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on facebook (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on linkedin (opens in a new window)
* Google’s Gemini hacked three companies in new AI safety incident on whatsapp (opens in a new window)
*

Stephen Morris in San Francisco

PublishedSeptember 19 2026

Jump to comments section

Unlock the Editor’s Digest for free

Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.