AI might not actually be killing off jobs like we thought - but many businesses are still wary about agents (www.techradar.com)

New research from Gartner shows enterprises are largely favouring supervised, human-in-the-loop AI over fully autonomous agents: only about 15% are considering, piloting or deploying fully autonomous agents versus roughly 75% engaging with agentic AI overall. Few IT leaders believe agents will outright replace human workers soon (only 7% strongly agree on replacement within two to four years); instead 12% expect agents to replace legacy workplace applications and reshape workflows. Major practical concerns are holding back wider agent adoption — 74% see AI as a new attack vector and 53% doubt it will deliver truly transformative gains, even if it can boost productivity. Those findings, paired with analysis from Yale, suggest the current wave of AI-driven change looks more like cautious augmentation than wholesale job displacement. While companies such as Lufthansa and Salesforce announced ~4,000 roles cut each, Yale’s work argues those moves look like cost control rather than evidence of broad labor-market disruption since ChatGPT’s launch. Implication for the AI/ML community: enterprise adoption will emphasize supervised models, governance, security, and measurable ROI; vendors who deliver secure, explainable, human-centered agent systems and clear productivity cases will find more traction than those promising fully autonomous replacements.