Google Faces Backlash Over Deceptive AI Voice Assistant Duplex
Google demos a highly realistic AI called Duplex that books appointments via phone, sparking serious ethical concerns about the tech giant's approach to artificial intelligence design and deception.
Google CEO Sundar Pichai impresses a developer crowd at the I/O conference with a demo of Duplex, an AI system that makes phone calls to book appointments on behalf of users. The voice assistant successfully navigates a hair salon booking and a restaurant inquiry by using distinctly human-sounding voices that include natural vocal cues like "ums" and "ahs." The realistic pantomime effectively convinces the human receptionists on the other end of the line that they are speaking to a real person.
Despite the enthusiastic applause from the home crowd, the demonstration raises major red flags regarding Google's approach to ethical AI development. Critics point out that the company treats ethics as an after-the-fact consideration rather than a core component of the engineering process. This delayed ethical scrutiny occurs just as Google prepares for a real-world rollout of the technology, highlighting a troubling trajectory for its artificial intelligence innovations.
Experts specifically criticize the foundational design of the Duplex experiments, noting that the core hypothesis revolves around whether the AI successfully deceives a human. Researchers argue that developers do not need to trick people into believing they are talking to a human just to provide a smooth and natural user experience. This focus on deception over genuine usability ultimately shows a significant flaw in how Google designs and tests its cutting-edge AI systems.