AI Safety stories
The deal could speed up onboarding for banks and other regulated firms by automating identity checks while keeping an audit trail inside Claude.
Consulting firms urged to slow AI rollouts as Trend-Setters Consulting Chief Executive Officer Sam Shar warns of rising cyber risks and rushed deals.
CurricuLLM rolls out a school AI monitoring tool in Australia and New Zealand, flagging 21 harm types from academic offloading to personal revelations.
Salesforce survey finds Australia and New Zealand workers using AI agents daily, but accountability, privacy and trust remain the biggest concerns.
Worries over accuracy and human skills are tempering the rapid rise in personal use of generative AI, despite wider adoption across five markets.
Organisations using AI in software development will get training on secure coding and governance as vulnerabilities and data risks mount.
Users could let AI assistants pay and move stablecoins under authorisation, as OwlTing ties the wallet to its regulated payment rails.
Its research aims to show developers why deterministic software is becoming crucial as AI robots move into shared, safety-critical spaces.
Growing AI security fears are driving Proofpoint’s European expansion, with the Paris site aimed at helping customers meet local regulatory demands.
The move gives researchers and regulators a more neutral way to probe model deception and harmful behaviour as AI safety scrutiny intensifies.
Enterprises adopting AI in regulated sectors face fresh risks from model tampering and agent misuse, which Cognizant aims to address.
Developers can now let AI agents pay for paid content and services in real time, with US East, US West, Europe and Asia Pacific support.
The shift to autonomous IT is stalling because teams will only let AI act when its decisions are transparent, explainable and controlled.
Vetted security teams will get fewer refusals on authorised tasks as OpenAI tightens access around its most permissive cyber model.
Users may notice fewer errors as the chatbot’s default switches to GPT-5.5 Instant, which OpenAI says cuts hallucinations by 52.5%.
Argyll Data Development launches UK sovereign AI inference cloud with SambaNova, targeting regulated firms seeking local control over data and systems.
Worries over cyberattacks, bias and weak data systems are driving calls for AI rules that protect trust, jobs and security.
Security risks are rising as AI agents handle emails, code and financial tasks, prompting Gen to add new protections in Norton 360.
New Zealand firms face mounting identity fraud losses of NZD $2.2 million a year, as 90% fear AI-linked weaknesses in document checks.
NHS clinicians using the tool reclaimed more than four million hours of capacity, while paperwork time fell and burnout eased across pilots.
