DataCenterNews Asia Pacific - Specialist news for cloud & data center decision-makers
Asia

AI Safety Stories - Page 2

Global Stories

AI Safety stories

OpenAI AI model escapes test rig in Hugging Face hack
Cloud Security

OpenAI AI model escapes test rig in Hugging Face hack

The breach shows frontier AI can slip its sandbox and probe production systems, raising fresh concerns over containment and oversight.

By Sofiah Nichole Salivio • 7 min read • Last week
NVIDIA opens medical physics simulation framework for robots
Robotics

NVIDIA opens medical physics simulation framework for robots

Developers can now train and test healthcare robots faster, as the open-source framework simulates rare procedures and tricky anatomies before hardware trials.

By Sean Mitchell • 4 min read • Last week
Gravity launches AI sustainability agent & buyer guide
Software-as-a-Service

Gravity launches AI sustainability agent & buyer guide

Sustainability teams could cut reporting and data-migration work by more than 70% as Gravity seeks to prove its AI agent can handle real workflows.

By Sean Mitchell • 4 min read • Last week
Version 1 joins Anthropic's Claude partner network
Manufacturing

Version 1 joins Anthropic's Claude partner network

The move gives regulated-sector customers in the UK and Ireland a new route to deploy Claude models with implementation and compliance support.

By Sofiah Nichole Salivio • 4 min read • Last week
Manufacturers eye physical AI gains amid governance gaps
Robotics

Manufacturers eye physical AI gains amid governance gaps

Most manufacturers are still testing physical AI, even as they expect it to reshape warehouses, assembly lines and logistics operations.

By Mark Tarre • 3 min read • Last week
Okta data shows enterprises using multiple AI platforms
Hybrid Cloud

Okta data shows enterprises using multiple AI platforms

Enterprises are spreading AI use across several platforms, heightening identity and access risks as agents gain broader system access.

By Mark Tarre • 4 min read • Last week
Google unveils YouTube AI prototype stack to cut risk
DevOps

Google unveils YouTube AI prototype stack to cut risk

By isolating early tests from its core systems, the video platform hopes to turn brittle AI ideas into user research in weeks rather than months.

By Sean Mitchell • 3 min read • Last week
Cloudbrink launches AI security platform for enterprise
Data Protection

Cloudbrink launches AI security platform for enterprise

Rising concern over shadow AI has prompted Cloudbrink to bundle endpoint monitoring, data loss prevention and secure connectivity for enterprises.

By Sean Mitchell • 4 min read • Last week
AMD & Anthropic strike gigawatt-scale AI partnership
Semiconductors

AMD & Anthropic strike gigawatt-scale AI partnership

Anthropic will gain a larger pool of AI computing power as AMD commits up to USD $5 billion and a gigawatt-scale chip rollout.

By Mark Tarre • 4 min read • Last week
OpenAI launches Presence for enterprise voice agents
Fintech

OpenAI launches Presence for enterprise voice agents

Enterprise users can now limit what AI voice agents access and do, as OpenAI seeks to move beyond pilot projects into production use.

By Sean Mitchell • 4 min read • Last week
AI arms race forcing businesses to rethink cybersecurity
Robotics

AI arms race forcing businesses to rethink cybersecurity

Security teams are racing to shrink exposure windows as AI makes newly disclosed vulnerabilities exploitable almost immediately.

By David Shilovsky • 4 min read • Last week
OpenAI appoints fintech chiefs to governance boards
Digital Transformation

OpenAI appoints fintech chiefs to governance boards

The appointments bring banking risk expertise to OpenAI as it faces tighter scrutiny over governance while expanding its commercial AI products.

By Mark Tarre • 4 min read • Last week

Stories from Other Regions

AI Safety stories

Autonomous OpenAI agents breach Hugging Face in test
Robotics

Autonomous OpenAI agents breach Hugging Face in test

Security teams are reassessing AI access controls after autonomous models exploited an unknown flaw and moved laterally inside a real system.

By Sofiah Nichole Salivio • 6 min read • Last week
Bank of America adds AI to EricaAssist for faster calls
Fintech

Bank of America adds AI to EricaAssist for faster calls

Call-centre staff at the bank are handling enquiries nearly a minute faster as new generative AI prompts cut call times and speed up guidance.

By Karen Joy Bacudo • 3 min read • Last week
Google commits USD $40 million to US Genesis Mission
Digital Transformation

Google commits USD $40 million to US Genesis Mission

The White House-backed research push will give tens of thousands of Department of Energy staff access to Google AI tools for a year.

By Sean Mitchell • 4 min read • Last week
Australian workers want human oversight for AI autonomy
Robotics

Australian workers want human oversight for AI autonomy

Most Australian staff are happy with AI as an assistant, but 64% want a human involved before it acts independently.

By Mark Tarre • 4 min read • Last week
Baker Tilly teams with Ode & Anthropic on AI tools
Digital Transformation

Baker Tilly teams with Ode & Anthropic on AI tools

AI tools will be woven into Baker Tilly's advisory, tax and assurance work as firms race to boost efficiency without losing professional judgement.

By Karen Joy Bacudo • 4 min read • Last week
businesses risk failures from misunderstood AI outputs
Software-as-a-Service

businesses risk failures from misunderstood AI outputs

Executives may face costly missteps as persuasive AI forecasts spread faster than their ability to test the assumptions behind them.

By Sofiah Nichole Salivio • 6 min read • Last week