EU AI Act checker reveals Big Tech's compliance pitfalls


A recent analysis indicates that major AI models from companies like OpenAI, Meta, and Alibaba are not fully compliant with upcoming EU regulations, particularly regarding cybersecurity resilience and discriminatory outputs. The evaluation was conducted using the "Large Language Model (LLM) Checker," a tool developed by Swiss startup LatticeFlow in collaboration with researchers. While models generally scored well, with averages above 0.75, significant deficiencies were found in areas like discriminatory output, where OpenAI's "GPT-3.5 Turbo" scored only 0.46. As the EU prepares to enforce the AI Act over the next two years, companies face potential fines for non-compliance. LatticeFlow aims to provide developers with a roadmap for aligning their models with these regulations.

Read More


thumb-image

Solutions