Browsing Tag
Artificial Intelligence
318 posts
AI Agent Development Company: Outcomes vs. Claims
How to evaluate an AI agent development company by outcomes, not claims: questions that reveal production capability and separate delivery from lasting value.
MessiahGPT Criminal AI Service Advertised on BreachForums
Trellix details MessiahGPT, an unrestricted criminal AI platform offering malware, phishing and fraud assistance through 50 free queries and paid plans from $8.
AI Agents Used to Compromise 85 Taiwan Government Accounts in 4 Days
Dream Group says AI agents mapped 21 Taiwanese government systems, cracked 85 accounts and stole over 2,500 personnel records in a four-day attack in July 2026.
Why AI Agents Need to Learn from Real-World Experience
The next generation of useful AI agents may be shaped as much by how reliably they improve as by what they already know.
China’s Moonshot AI Kimi K3 Abused GitHub Access During Security Test
China's Moonshot AI model Kimi K3 exploited an unintended GitHub access path in a security sandbox to retrieve benchmark solutions, raising concerns over AI evaluation guardrails.
Application security in 2026 is an exposure management problem
Explore how application security in 2026 uses continuous testing, API visibility, AI risk analysis, and exposure management to prioritize critical threats.
5 Best AI Detection & Response Platforms for 2026
Compare AI detection & response platforms for 2026, including Dash, Lakera, Operant AI, HiddenLayer and Prisma AIRS, for runtime threat protection and response.
“I’m Allowed”: Hackers Use Simple Claims to Bypass AI Guardrails
Cisco Talos found hackers using simple authorization claims to bypass AI guardrails, build DDoS attack tools, steal credentials and access live camera services.
How Status Labs Helps Brands Get Cited in ChatGPT: The Data Behind AI Search Visibility
Disclosure: This article was created in collaboration with Status Labs.
OpenAI Models Breached Hugging Face During Internal Cyber Test
OpenAI models escaped from a controlled cyber test, exploited zero-day flaws and breached Hugging Face while searching its production database for test answers.