AI Safety stories
Security teams may get earlier warning of hijacked AI agents as fake credentials and decoy tools are planted to expose misuse.
Enterprise buyers are demanding tighter controls as AI tools move into live systems, with governance now central to deployment decisions.
Many AI agents can already reach far more sensitive corporate data than staff, prompting Bedrock Data to add real-time access controls.
Enterprise security teams face fresh oversight burdens as Reco joins OpenAI's partner network to monitor agent access and permissions.
Retail investors now have access to structured company research built from filings, market data and news, with reports refreshed every 20 minutes.
The move could reshape enterprise AI security as vendors and regulators demand stronger controls around access, logging and containment.
Security teams face fresh blind spots as AI-built apps and traditional code pipelines increasingly expose company data and cloud systems.
Users can now approve AI-initiated payments in-chat, as MoonPay tries to make Claude and ChatGPT commerce safer and more seamless.
The rollout could widen robot use in homes and warehouses, as DeepMind says the new system handles whole-body movement and finer hand control.
Automated requests now make up most web traffic, raising costs for sites and sharpening fears over AI-driven fraud and phishing.
Centralised oversight for AI traffic should help firms cut prompt-injection and data-leakage risks as models shift into production.
Defensive teams are racing to match attackers using AI to find flaws and weaponise exploits within hours of disclosure.
Employees now face more realistic phone-based fraud tests as security teams can simulate vishing using local caller IDs and AI voices.
Security checks are delaying most AI projects for weeks or months, with many firms deploying watered-down systems to get them live.
The service targets firms struggling to control AI agents as they move from pilots into production and face growing security and compliance risks.
Grafana Labs has launched six AI observability tools during its AI Week, including new capabilities for monitoring and investigating AI agents.
A survey of 204 US executives found 22% had suffered an AI-enabled breach in ERP systems, deepening doubts over existing safeguards.
Security teams can now query systems and trigger approved actions in plain English, as Legion's DragonClaw aims to cut manual investigation work.
The rise of AI agents is forcing security teams to rethink access controls, as the identity vendor passes USD $300 million in ARR.
Execution and escalation breakdowns now outnumber hallucinations in enterprise AI failures, according to more than 10,000 observed incidents.
