OpenAI stories
Finance teams can now track AI costs across providers, with Ramp saying its tool can flag waste and cut monthly spend by 12%.
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.
Regulated firms can let non-technical staff build apps in a controlled browser, using approved AI tools and existing security controls.
Energy and cooling limits are becoming the real bottleneck for AI operators, as NVIDIA says Blackwell racks can lift output within fixed power budgets.
Rising usage-based bills could now be easier to control, as 1Password's SaaS Manager adds dashboards, limits and alerts for AI spend.
Corporate training teams could cut admin time as the Dublin software provider adds AI agents and external tools to its learning platform.
The deal values the web data infrastructure group at USD $3.6 billion as investors back the systems helping AI agents gather online information.
Plaud has launched a ChatGPT app that lets users search, summarise and analyse meeting recordings stored in their Plaud accounts.
Users can now pull Dropbox files into ChatGPT, Claude and Gemini Spark, keeping AI output tied to existing permissions and team workflows.
Users who miss the deadline will lose access to OpenAI's most advanced cyber models, as the firm tightens defences against phishing.
Microsoft 365 users may see stronger first drafts and faster analysis as OpenAI's latest model becomes the default Copilot engine.
Measured results are now under pressure as firms struggle to turn AI pilots into everyday tools and prove climate plans beyond targets.
By tying strategy to revenue outcomes, the platform aims to help mid-sized firms prove marketing's return and cut tool sprawl.
Public servants tested ChatGPT and Codex on drafting and research, as agencies weigh whether AI can speed up routine work without losing human oversight.