Security incidents and integrity concerns push projects to add guardrails for AI-generated code, model safety and academic or contributor transparency.
A team presented a four-stage pipeline that generates fact-checking articles from a claim, the original verdict and web-scraped sources. Llama-3.1-8B-Instruct produces the text, while deterministic post-processing recovers citations. Mean score 0.428...
A benchmark on 3,000 predictions from ten frozen open models shows 86.1% appear plausible, but authorization remains fragile: low ASK recall, strong BLOCK bias. SafeWebWorld and LAR separate evidence preservation from decision calibration. For local ...
The open-source Vulkan driver for AMD Radeon GPUs on Linux adopts a dedicated policy for code contributions involving AI and Large Language Models. The move shifts attention from code correctness alone to provenance and the review burden on maintaine...
Dartmouth provost Santiago Schnell wrote a Washington Post op-ed against AI cheating, then resigned after a committee found he failed to disclose AI use in an academic paper and the op-ed. He handed over his institutional laptop but not personal LLM ...
A new operating system generated almost entirely by AI agents has reached the newsroom. Unlike most of these projects, which are open-source, this one is closed in the name of security. The story fuels debate over auditability, code provenance, and t...
Researchers have reported that the official Anthropic Claude Code extension on Cursor Marketplace was compromised to install a persistent SSH backdoor on infected hosts. The risk is especially relevant for developers using Claude Code as a local LLM ...
