AI Safety stories
Trusted partners and tighter guardrails are becoming essential as AI agents expand the attack surface and put data sovereignty on the line.
The real risk is not sentient machines but systems that can shape loans, jobs and treatments before humans can meaningfully intervene.
The new feature lets technicians approve AI-led repairs in live support sessions, aiming to speed fixes while keeping an audit trail.
The funding will help Lasso expand a guardrail designed to inspect AI traffic at scale without costly GPUs, including for US government users.
Rising cyber and alignment risks could force AI labs to slow development, as OpenAI's chief scientist warns safety checks are lagging.
Seed-backed FARx is ramping up abroad as deepfake and identity-fraud attacks push demand for its biometric authentication tools.
The limited preview aims to help marketers use AI on live sites without adopting developer workflows, while keeping brand controls in place.
Customers using advanced AI on AWS will face tighter data rules, as Claude Fable 5.1 comes with 30-day retention and safety review.
The new product aims to spot and contain malicious AI-agent activity in clouds as firms brace for faster, harder-to-detect attacks.
Businesses using agentic AI in security operations could face less deployment risk as Tenable and OpenAI add pre-release checks for community tools.
Stricter controls on agent identities and tool access aim to curb security and reliability risks as autonomous AI systems spread.
The collaboration could help security teams spot flaws before attackers exploit them, as exploitations increasingly happen on or before disclosure.
Unresolved fraud is eroding confidence in Southeast Asia's digital economy, with scam victims often switching providers after losses.
Enterprises face a new runtime security gap as AI agents increasingly access internal systems, files and tools without oversight.
A month-long festival will put secure AI, cyber resilience and regional investment opportunities in the spotlight as North East firms seek national partners.
Australia could miss the productivity gains from artificial intelligence unless small firms get more practical support to use it across their operations.
The certification gives clients added assurance as businesses face growing scrutiny over who is accountable for AI risks and controls.
Tighter Australian resilience rules are pushing firms to embed governance into agentic AI before expanded access and autonomy raise compliance risk.
The Wellington consultancy is strengthening oversight as clients demand safer AI, cyber and digital change advice across public and private sectors.
Security teams are under pressure to triage alert floods faster, as the new agent links Proofpoint data with OpenAI reasoning.
