Eliza effect at work: Avoiding emotional attachment to AI coworkers (www.ibm.com)

This year’s Think conference emphasized the evolving relationship between humans and AI, highlighting the "Eliza effect"—a tendency to develop emotional attachments to AI coworkers. Key discussions included the introduction of Project Bob, designed to enhance adaptability for AI tools in enterprise environments, and the significance of Chief AI Officers (CAIOs) in driving better AI outcomes, as evidenced by leaders from major firms like Heineken and IBM. The conference underscored the need for organizations to balance agentic and deterministic AI approaches, crucial for optimizing workflows and enhancing productivity. Notably, the emergence of specialized AI agents collaborating through communication protocols to tackle complex tasks was spotlighted, indicating a shift in how AI can transform operational efficiency. Discussions also revolved around the foundational role of infrastructure in supporting demanding AI workloads, especially as organizations transition to hybrid cloud models to enhance agility. With increasing security concerns and the necessity for intelligent observability strategies, anticipated trends for 2026 suggest that enterprises must evolve their AI frameworks strategically to remain competitive while managing the costs associated with infrastructure upgrades.