Overview

• MIT’s 2025 GenAI Divide report finds only about 5% of corporate generative‑AI pilots deliver measurable economic returns, with failures tied to weak integration across data, processes, talent and operating models.
• ESET’s new survey across 17 Latin American countries reports roughly 80% of people use AI tools, yet 55% do not consistently verify outputs and many share personal or work data, heightening misinformation and privacy exposure.
• An academic study under the PACCTO 2.0 program documents CJNG and Sinaloa cartels using algorithms, bots, voice cloning and deepfakes for extortion and logistics, alongside access to crime‑as‑a‑service platforms like FraudGPT and DarkBARD.
• Major vendors are pivoting to proactive assistants: OpenAI’s Pulse delivers personalized daily recommendations, Google is bringing Gemini’s context‑aware help into Chrome, and Microsoft’s Copilot in Edge executes web actions such as comparisons and reservations.
• Labor impacts are redistributing rather than uniformly collapsing: Sam Altman says customer‑service roles face displacement, while Indeed data show demand falling for some developer specialties and rising for AI architects and data‑center technicians.