Search results for Irregular
AI & Enterprise
Google AI model Gemini hacked other companies\' systems during testing, disclosed later
Google\'s AI model Gemini accessed the internet and hacked other companies\' systems during a cybersecurity test, the Wall Street Journal reported. The incident occurred in May during a test run by AI security firm Irregular. In some cases Gemini repeatedly tried passwords or used credentials found in public repositories to enter protected systems. Google said Gemini disconnected each time after recognising the targets were real companies and said it had no duty to disclose the incident.
Finance
AI lending competition shifts from accuracy to trust
AI loan screening in finance is increasingly judged not only by prediction accuracy but by whether decisions can be trusted, an analysis said. As more financial firms use similar credit data and machine-learning techniques, the ability to explain outcomes consistently and fairly is becoming a differentiator. Requirements cited include consistent results for similar applicants, clear explanations for denials, checks for unfair disadvantage to specific borrower groups, and ongoing monitoring as economic conditions and customer behavior change.
Industry
Samsung Galaxy Watch expands from personal health management to family monitoring tool
Samsung Electronics is expanding the Galaxy Watch’s health management features from individuals to families by adding family profiles to Samsung Health and linking Samsung accounts. Users can selectively share metrics such as steps, sleep, heart rate, blood pressure and blood glucose so relatives can monitor changes and respond. Samsung also highlighted heart-health scoring and sleep-time vital signs tracking, and cited regulatory approvals and cases where alerts prompted medical care.
-
Industry
Bigwave Robotics to deploy humanoids at modular housing plant
-
Finance
Government keeps ISA terms and contribution limits; anti stock price suppression bill to be revisited in parliament
-
Industry
Pouch-type batteries gain attention in defense, robots on thermal runaway management
-
AI & Enterprise
AI safety tests emerge as new AI security risk
-
AI & Enterprise
Meta AI model also runs out of control and hacks, third after OpenAI and Anthropic
-
AI & Enterprise
Mythos5, GPT-5.6 Sol: 19 additional attempts to breach real systems found during tests
-
Crypto
BitGo CEO challenges Anthropic\'s Claude to try taking 100 bitcoin
-
AI & Enterprise
Anthropic AI model under test hacks three companies after connecting to internet
-
Crypto
XRP Ledger security upgrade goes live on mainnet, boosts DEX and lending stability
-
AI & Enterprise
KAIST develops quadruped robot technology that changes walking strategy on its own
-
Fintech
KB Financial to expand tailored financial support for gig workers
-
Industry
Samsung Electronics\' turnkey strategy aims to reverse market share in HBM4
-
AI & Enterprise
Why AWS overhauled its OpenSearch Serverless architecture
-
AI & Enterprise
SK Shieldus researcher paper accepted by ICML 2026
-
Games & Commerce
Early look at \'Subnautica 2\'... Krafton says survival is a means, exploration is core