Talking to an AI assistant is becoming increasingly similar to having a natural conversation. Google has now introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two new AI models designed to make real-time voice interactions more intelligent, responsive and useful.
Google says the new models improve near-real-time reasoning and are designed to handle more complex conversations and multi-step tasks through voice.
What is Gemini 3.8 Live?
Gemini 3.8 Live is Google's latest voice-focused AI model. It is designed to respond naturally during live conversations instead of requiring users to type every question.
The model combines conversational intelligence with visual grounding, allowing Gemini to work with information beyond just the words being spoken.
This could make voice assistants more useful for situations where typing on a phone is inconvenient.
A second version is designed for harder tasks
Google has also introduced Gemini 3.8 Live Extended Thinking.
As its name suggests, this version is designed for more complex problems that require additional reasoning and multiple steps.
Instead of simply answering a short question immediately, the model is designed to spend more effort working through complicated requests.
For example, users could potentially use an AI assistant to break down a complicated task, compare different pieces of information or work through a multi-step problem using voice.
Conversations are becoming more natural
One of the biggest changes in modern AI assistants is the move away from traditional question-and-answer interactions.
Earlier voice assistants were generally designed around short commands such as setting an alarm, checking the weather or playing music.
Newer AI models are being developed to maintain a conversation, understand context and respond to follow-up questions without requiring users to repeat everything.
Gemini 3.8 Live is part of this broader shift toward more conversational AI.
AI can also understand visual information
Google says Gemini 3.8 Live includes visual grounding capabilities.
This means voice interaction can be combined with visual information, potentially allowing users to discuss something they are showing to the AI rather than describing everything manually.
This could be useful for activities such as understanding an object, discussing a document or asking questions about something visible through a compatible camera experience.
What is Extended Thinking useful for?
The Extended Thinking version is aimed at situations where the answer requires more than a quick response.
Complex planning, detailed reasoning and multi-step problem solving are examples of tasks where additional reasoning could be useful.
The important difference is that the AI is not simply being designed to respond faster. It is also being designed to spend more computational effort when a problem requires it.
Could this change how people use AI?
The development of models such as Gemini 3.8 Live suggests that AI assistants are moving beyond simple chatbots.
Instead of opening an app, typing a question and waiting for an answer, users can increasingly interact with AI through a continuous conversation.
This could make AI more useful while driving, working, studying or performing other tasks where typing is inconvenient.
However, users should still verify important information generated by AI, particularly when making financial, legal, medical or other high-stakes decisions.
The next phase of AI is becoming more conversational
Google's latest Gemini models show how AI companies are focusing not only on making models more intelligent but also on making interactions feel more natural.
With real-time voice conversations, visual understanding and extended reasoning being combined into a single experience, AI assistants are gradually becoming tools that users can interact with more like digital collaborators than traditional search boxes.
