Explore the intersection of AI and cybersecurity. Stay informed on AI-driven security trends, tools, and threats in the ever-evolving digital landscape.
Search across headline titles and summaries.
Background for this topic.
Artificial intelligence (AI) describes computer systems that perform tasks such as recognizing patterns, making predictions, understanding language, or generating content. In security reporting, the term commonly includes machine-learning models used for detection and analysis, as well as generative AI applications that produce text, code, images, or other outputs.
AI can help analyze security telemetry, prioritize vulnerabilities, and support investigations, but its outputs can be wrong or manipulated. Important attack surfaces include prompt injection that steers an application into unintended actions, sensitive data being exposed through prompts or model outputs, and excessive permissions granted to AI systems that use external tools. Models can also be degraded by poisoned training data or evaded with carefully crafted inputs. Practitioners should protect training and operational data, limit model access and tool permissions, test for adversarial behavior, and require appropriate human validation before high-impact decisions.
Cybersecurity researchers have disclosed a zero-click flaw in OpenAI ChatGPT's Deep Research agent that could allow an attacker to leak sensitive Gmail inbox data with a single crafted email without any user action
Startup Simulates Offensive and Defensive AI to Test and Thwart AI-Based ThreatsIrregular secured $80 million in funding to turn its research into scalable security tools for businesses adopting AI. With growing offensive AI capabilities, the company is racing to productize simulations that detect vulnerabilities before attackers do.
Funding Supports Threat Hunting, Natural Language to Replace Legacy DetectionVega aims to replace patchwork AI integrations with an analytics layer that enables real-time, natural language detection across distributed data. Backed by Accel, the company will double headcount, improve detection tuning and reduce false positives without a SIEM rip-and-replace required.
Zero Trust could help organizations fight back against attackers who use artificial intelligence, but new threats will require the architecture to evolve.
The loophole allows cyberattackers to exfiltrate company data via OpenAI's infrastructure, leaving no trace at all on enterprise systems.
Microsoft has begun rolling out the beta version of its AI-powered Gaming Copilot to Windows 11 systems for users aged 18 or older, excluding those in mainland China. [...]
Also: Inside the AI Sandbox Act; Cybersecurity Summit: London Financial ServicesIn this week's update, three ISMG editors discussed Mitsubishi Electric's $883 million purchase plans for Nozomi Networks, the new Sandbox AI bill in Congress that aims to cut perceived red tape limiting AI innovation, and takeaways from the latest Cybersecurity Summit: London Financial Services.
Could this bot-prevention technique now be obsolete? ChatGPT can be tricked via cleverly worded prompts to violate its own policies and solve CAPTCHA puzzles, potentially making this human-proving security mechanism obsolete, researchers say.…
New Texas health information legislation that began to go into effect on Sept. 1 includes several noteworthy provisions including requirements related to health record data storage and artificial intelligence, said regulatory attorney Rachel Rose. Rose explains the significance of the new state law.
Researchers at Radware found a zero-click flaw in ChatGPT Deep Research agent when connected to Gmail and browsing
Run by the team at workflow orchestration and AI platform Tines, the Tines library features over 1,000 pre-built workflows shared by security practitioners from across the community - all free to import and deploy through the platform's Community Edition
Radware says flaw enabled hidden email prompts to trick Deep Research agent into exfiltrating sensitive data ChatGPT's research assistant sprung a leak – since patched – that let attackers steal Gmail secrets with just a single carefully crafted email.…
Trend Micro said the use of AI platforms to create and host fake CAPTCHA pages helps attackers develop more sophisticated phishing campaigns at scale and speed
.OpenAI released $4 GPT Go in August, but it was limited to just India. Now, OpenAI is expanding GPT Go to include new regions. [...]
Cybercriminals are abusing AI-native platforms like Vercel, Netlify, and Lovable to host fake captcha pages that deceive users, bypass detection, and drive phishing campaigns.
OpenAI has rolled out a big update to ChatGPT Search, which is an AI-powered search feature, similar to Google AI Mode. [...]
Jennifer Ewbank, former CIA deputy director of digital innovation, on resilience, cultural shifts, and cyber fundamentals in the AI era.
Silicon Valley Startup Brings AI Agent and Prompt Injection Protections to FalconCrowdStrike plans to purchase Pangea to add native AI detection and response capabilities to its Falcon platform. The company says the acquisition will help secure AI models and users alike from preventing prompt injection to tracking agent activity across enterprise environments.
OpenAI is finally rolling out a toggle that allows you to decide how hard the GPT-5-thinking model can think. This feature is rolling out to Plus and Pro subscribers. [...]
Also, Colt Services Outage Persists, Finland Charges Americans in Vastaamo HackThis week, Microsoft hit RaccoonO365, Colt Technology Services, Finland charged a U.S. citizen in Vastaamo hack. RevengeHotels hackers used AI, Meta can't overturn a privacy case verdict. Chinese hackers unleashed spear phishing emails. Prosper confirmed a data breach, as did Kering fashion houses.