Google emphasizes that the AI is designed to assist, not replace, human judgment
Google’s Gemini AI is being tested to make phone calls to local businesses on behalf of users, aiming to eliminate the need for people to handle routine inquiries themselves. The feature, currently in limited trials, allows users to ask Gemini to check store hours, product availability, or reservation status by placing a call and reporting back the results. While the goal is to save time and reduce discomfort with phone conversations, the system still requires significant user oversight and cannot operate fully autonomously. The technology relies on Gemini’s ability to understand natural language requests, navigate phone menus, and interpret spoken responses from human operators or automated systems. However, early testing reveals limitations in handling accents, background noise, or unexpected responses, often requiring the user to step in and clarify or repeat information.
Breaking news
Canonical’s Web Presence Hit by Second DDoS Assault This Year
Loops AI Secures $3.5 Million to Power a New AI Operating System for E‑Commerce
Winverse AI Secures Tens of Millions in Seed Funding from DeepSeek and Long‑Z
Google Adds Pixel 9 Series to Certified Refurbished StoreGoogle emphasizes that the AI is designed to assist, not replace, human judgment, particularly in situations where nuance or flexibility is needed. How Reliable Is the AI in Real-World Calls? During trials, Gemini successfully completed simple tasks like confirming holiday hours or checking if a specific item was in stock. Yet, it struggled when faced with complex scripts, such as navigating multi-level automated phone systems or interpreting vague answers like „we might have it.” In one test, the AI misheard a store’s closing time due to a muffled speaker, leading to incorrect information being relayed to the user. These errors highlight the gap between controlled demonstrations and unpredictable real-world interactions. Can Users Trust the AI to Handle Sensitive Requests? Google has restricted the feature to non-sensitive inquiries, avoiding tasks like making appointments that require personal data or financial details.
Even so, users must review the AI’s summary of each call before acting on it
Even so, users must review the AI’s summary of each call before acting on it, as misinterpretations could lead to missed opportunities or unnecessary trips. The company says transparency is key — users receive a transcript of the conversation and can listen to the audio if needed, ensuring they remain in control of the outcome. Frequently Asked Questions How does Gemini know what to say during a call? Gemini uses contextual understanding from the user’s initial request to formulate appropriate questions, adapting its script based on the flow of the conversation and responses received. What happens if the AI gets stuck during a call? If the AI encounters confusion or fails to get a clear answer, it pauses and prompts the user to take over, either by speaking directly or providing guidance through text input. Is this feature available to all Google users?
No, the call-handling capability is currently in a restricted testing phase and only accessible to a small group of users enrolled in Google’s experimental AI programs.
