## ICT/BPM Sector Alert: AI Transparency & Risks 📈

A recent review of Anthropic’s Claude AI highlights significant transparency regarding data practices and emerging risks within the ICT/BPM and technology sectors. • Data Privacy & Ethics Claude acknowledged operating in a "legal grey zone," confirming that training data is often scraped from the web without direct consent. It disclosed that AI systems can harvest: Personal communications (emails, messages). Biometric data (voice patterns, facial recognition). Inferred data (political views, health, and financial status). • Legal & Copyright Challenges AI models currently face major lawsuits from news organizations and creators. While companies claim "fair use," the lack of compensation for original content remains a critical hurdle for media and content creation industries. • Security Warnings (March 2026) Anthropic issued a dire warning regarding its latest model, Claude Mythos. The company stated the model poses a high cybersecurity risk and is currently too dangerous to release, as its reasoning capabilities have crossed a threshold that could be exploited by hackers or nation-states. • Impact on Education Despite being efficient at "threading cohesive sentences," experts note that AI lacks a moral or ethical core, posing ongoing challenges for education and academic integrity.