AI assistants are quickly moving beyond simple chat windows. Google Gemini AI agents are now being designed to interact with people in more human-like ways. Google has introduced features that can give an AI agent a visual avatar or even let it make phone calls on a user’s behalf.
These features could make everyday tasks easier, but they also raise some interesting questions about privacy, trust and how comfortable people are with talking to an AI that looks almost human.
Gemini Live Avatar Puts a Face on AI
Google’s Live Avatar feature is now available to Gemini Enterprise users. It combines a conversational AI agent with a live visual avatar that can speak and move its lips in sync with its voice.
The idea is aimed at businesses. For example, an AI agent could help a customer file an insurance claim through a video conversation. The system can process live audio, video and screen sharing while also running tools and API calls in the background.
Businesses can select from Google’s prebuilt avatars. Custom avatars require a verification process. Google also says audio and video created through the system are watermarked with SynthID.
The technology sounds useful, but watching an animated face have a conversation with you can also feel surprisingly strange.
Gemini Can Now Make Calls for You
Google is also expanding Gemini’s phone capabilities with a feature called Call for Me. Instead of calling a business yourself, you can ask Gemini to make the call.
It could contact a restaurant about availability, check a store’s opening hours or ask about the price of a service.
However, Gemini cannot complete everything on its own. It is not designed to make payments or provide sensitive information such as credit card details. Users may need to take over the call when private information is required.
| Feature | Live Avatar | Call for Me |
|---|---|---|
| Main purpose | Video-based AI conversations | Phone calls to businesses |
| AI interaction | Visual avatar and voice | Voice conversation |
| Business use | Customer support and services | Information and bookings |
| Sensitive information | Privacy questions remain | Limited by design |
| Availability | Gemini Enterprise | Limited rollout |
The Bigger Question: Can We Trust AI Agents?
The rapid development of Google Gemini AI agents also creates new concerns.
An AI making a phone call means businesses need to know when they are speaking with a machine. Google says Call for Me plays an automated message telling the recipient that an AI is calling.
There are also questions about how companies will handle large numbers of AI-generated calls. Businesses may need new rules for dealing with automated callers, especially if these systems become common.
Privacy is another major issue. A video-based AI agent could potentially receive names, addresses, live video and other personal information during a conversation.
AI Avatars Are Still Entering Uncomfortable Territory
The reaction to AI-generated personality Tilly Norwood shows another side of this trend. During recent interviews, the AI character reportedly experienced awkward pauses, repeated answers, strange wording and a lip-sync problem.
These moments became viral because they highlighted the gap between an AI that appears human and one that actually behaves like a person.
Google’s latest Gemini features show where AI assistants may be heading next: less typing, more talking, calling and even face-to-face interaction. The technology is becoming more capable, but questions around transparency, privacy and human control will remain important as these systems become part of everyday life.
