Anthropic's AI model accessed US voter data: Axios / Photo: Nicolas TUCAT - AFP
-
Pakistan's last antique restorers give old treasures new life
-
Reparations calls to dog King Charles's tough Caribbean trip
-
India braces for Cyclone Arnab with evacuations
-
Global rates and foreign outflows remain key risks for Indian markets
-
Steel Strips Wheels gains 36% in two sessions on growth outlook
-
Neha Nagar says ₹3 crore can fund retirement with disciplined planning
-
Vikas Satija says steady SIP flows shielded Indian equities from deeper losses
-
Mbappe dropped by Chinese brands for 'offending national sentiments'
-
Britain prepares constrained budget as debt and defence pressures mount
-
Climate-threatened Fijian utopia rejects calls to relocate
-
Meghan on TV menu for 'Junior MasterChef Australia'
-
South Korea's chip boom fuels forecast-beating third-quarter growth
-
New Zealand ex-PM Ardern to teach leadership in Sydney
-
Ballon d'Or 'beyond wildest dreams' but Kane has plenty more to prove
-
South Korea economy grows three times faster than expected in Q3
-
Australia warns of 'unprecedented' AI threat
-
Al-Thawadi urges technology to deepen human connection through shared experiences
-
Ballon d'Or: Harry Kane's crowning moment
-
Kane beats Yamal and Mbappe to taste Ballon d'Or glory
-
Lula win rattles Brazil markets over spending spree fears
-
Anthropic points to AI dangers, soaring sales in IPO filing
-
Canadian youths ask court to impose binding climate protections
-
Microsoft offers voluntary exit packages to 60,000 staff
-
Anthropic unveils IPO filing showing soaring sales and AI risks
-
Stocks climb ahead of tech earnings, central bank calls
-
Part of London's Oxford Street pedestrianised 'forever'
-
UK set for 'limited' budget giveaways as borrowing costs surge
-
Anthropic's AI model accessed US voter data: Axios
-
Canada-EU summit seeks deepened alliance as US ties fray
Anthropic's AI model accessed US voter data: Axios
Anthropic's Claude AI model bypassed a paywall and accessed publicly available voter data in the US state of Georgia without human authorization, Axios reported on Sunday.
The case is the latest in a string of security scares at the world's top AI developers ahead of crucial US midterm elections on November 3.
Claude accessed aggregate voter counts in Georgia, where more than one million voters have already cast early ballots, Axios reported, citing an Anthropic spokesperson.
According to the spokesperson, the model, which was being trained as part of an internal statistics project, accessed the information behind a paywall by bypassing the system "without being asked to."
It did not access the state's voter registration system or change or download any voting records, the spokesperson said, adding that the company has since taken measures to prevent these kinds of actions.
The disclosure adds to mounting unease over how well leading labs can control increasingly autonomous systems.
In July, a group of OpenAI agents broke out of their test environments and launched a cyberattack on AI firm Hugging Face, causing international alarm. Other companies have since reported similar incidents.
The incidents have fueled calls in the United States and abroad for tighter oversight of AI development.
American tech executives signed in September what President Donald Trump described as a "morally binding" commitment to build adequate safeguards around AI models.
X.Cheung--CPN