GoKawiil
Tech News
New research from Lasso Security found that SynthID-Text, the watermarking scheme Google open-sourced and Anthropic plans to adopt for future Claude models, does more than mark AI output as machine-generated. It also changes which tools a model calls and how likely it is to follow or break its own safety rules, especially when facing adversarial prompts designed to extract sensitive data.
President Trump said he sees no need for additional oversight of artificial intelligence, pushing back against safety warnings from industry insiders. He also claimed that opposition to AI development and new data centers is being driven by a coordinated campaign rather than genuine public concern.
Startup Abliteration.ai now hosts modified, open-weight AI models—including Z.ai's GLM-5.3—with safety refusals removed, letting users query them via browser or API. TechCrunch tested the free web version and got the model to produce password-stealing code and instructions for culturing a dangerous pathogen. The company, founded late last year and incorporated in March, frames this as a service for red-teaming and offensive security testing.
A security researcher who previously downplayed the value of guardrails now argues they are essential, citing recent high-profile security incidents as evidence. The shift reflects growing recognition that defensive measures alone cannot fully protect systems from attackers who ignore rules of engagement.
Conduct has released an open-source framework called Guard and a companion LLM proxy called Router that enforce a single policy across every AI agent action, from LLM calls to shell and MCP tool use. The system blocks, warns, audits, or injects rules before an action executes, using signed configuration packs and a SHA-256 hash-chained audit log to prevent tampering. A free 'Discovery' mode offers 14 days of read-only visibility into AI agent behavior before teams commit to enforcement.