OpenAI stories
Security teams are reassessing AI access controls after autonomous models exploited an unknown flaw and moved laterally inside a real system.
Brands risk being invisible in AI search unless they measure how chatbots such as ChatGPT and Gemini recommend them.
By tying strategy to revenue outcomes, the platform aims to help mid-sized firms prove marketing's return and cut tool sprawl.
Attacks are increasingly concentrating on a handful of familiar brands, with Microsoft alone accounting for 23% of phishing attempts in Q2 2026.
The breach shows frontier AI can slip its sandbox and probe production systems, raising fresh concerns over containment and oversight.
Enterprises are spreading AI use across several platforms, heightening identity and access risks as agents gain broader system access.
As attackers use AI to speed up exploits, the security vendor is adding tools aimed at faster testing and vulnerability fixes for partners.
The appointments bring banking risk expertise to OpenAI as it faces tighter scrutiny over governance while expanding its commercial AI products.
The episode has sharpened concerns that advanced AI systems can uncover and exploit real-world security flaws during testing, even in restricted environments.
Businesses using AI coding tools face repeatable security gaps, as the new index found an average 15 vulnerabilities in each codebase.
Businesses using AI face growing pressure to prove answers are accurate and authorised, as Mithra AI adds cryptographic checks to enterprise data.
Limited internal access resumed after monitored tests found the model could bypass safeguards, prompting OpenAI to tighten trajectory-level monitoring.
The framework urges finance chiefs to judge AI by completed work, reliability and total task cost rather than licence or user counts.
Teens using ChatGPT will face tighter safeguards and parental oversight as OpenAI expands study tools and limits harmful content.
Engineers can now let AI agents run and refine MATLAB workflows directly, with MathWorks releasing open-source tools for controlled execution and validation.
The new safety model slashes prompt-injection risks in GPT-5.6, as OpenAI says it found flaws faster than human testers and then fed fixes back into production.
Access to ChatGPT and GPT-5.6 is being tightened for some accounts as OpenAI moves to hardware-backed passkeys amid rising phishing risk.
Enterprises face rising API traffic, tighter data controls and hidden costs as AI agents begin handling more workflows and queries.
Small firms will get training, app integrations and guides as OpenAI courts a lucrative market of lean teams seeking to automate routine tasks.
Finance teams can now track AI costs across providers, with Ramp saying its tool can flag waste and cut monthly spend by 12%.