Humans are still doing the work behind some AI agents (www.cnbc.com)

Recent developments in AI indicate that many personal AI agents still rely on human intervention for tasks that require nuanced decision-making, such as making phone calls. Companies like Dazzle AI, Duckbill, and Wajo AI emphasize the importance of human workers to bridge the gap in capabilities of foundation models developed by Meta and OpenAI. While these AI frameworks excel at surfacing information and making basic recommendations, they can struggle with complex tasks, such as negotiating service contracts or scheduling appointments. In many instances, human contractors step in to complete what the AI agents cannot handle effectively, ensuring that users receive reliable service. This hybrid approach, termed "collaborative AI," is significant for the tech community as it highlights both the potential and limitations of current AI technology. As AI agents encounter roadblocks, particularly with real-world judgment calls, companies are adapting by integrating APIs that connect these agents to human workers for complex tasks. This model not only enhances effectiveness but also raises essential discussions about privacy and transparency, especially as consumer trust in AI continues to develop. As foundation models improve, the AI industry is in a transition period where human support is still a crucial component of delivering quality service.