How ChatGPT’s New Voice and Image Features Transform AI Interaction

21CTO Guide: ChatGPT, the wildly popular AI chatbot, has always been just a text box. Now it’s learning to understand people’s questions in new ways.

Background

OpenAI is further refining the actions that AI‑driven chatbots like ChatGPT can perform, including the types of questions they can answer, the information they can access, and improvements to the underlying models.

This time the company has changed how users interact with ChatGPT, launching a new service version that allows prompts not only by typing sentences but also by speaking aloud or uploading images.

OpenAI’s Move

On September 25, OpenAI announced new voice and image features for the popular conversational AI robot ChatGPT.

These features represent a major expansion, enabling users to converse with the AI assistant and show it images for more natural dialogue.

OpenAI stated: “We are rolling out new voice and image capabilities in ChatGPT. They provide a more intuitive interface that lets you have spoken conversations or show ChatGPT what you’re talking about.”

Detailed Description of the Multimodal ChatGPT Add‑Ons

The new voice feature lets users interact with ChatGPT by speaking aloud. Users can choose from five AI‑generated voices, then ask questions or give instructions.

OpenAI prompts users: “Talk to ChatGPT and let it respond. You can chat anytime, ask it to tell a bedtime story for your family, or settle a table‑side dispute.”

Users feel as if they are talking to Apple’s Siri, Amazon’s Alexa, or Google Assistant, but with OpenAI’s underlying technology improvements, the answers are more precise. Large language models are reshaping virtual assistants, and OpenAI is leading the way.

The image feature allows users to upload photos to ChatGPT for visual information or queries. For example, a user can show a picture of their fridge and ask for recipe ideas, or send a landmark photo while traveling to have a real‑time conversation about it. The mobile app also includes a drawing tool that lets the AI focus on specific image regions.

OpenAI says these new capabilities are powered by its latest natural‑language AI models, GPT‑3.5 and GPT‑4, which can apply reasoning to visual and audio inputs. ChatGPT can now respond with one of five synthetic voices.

The company plans to roll out voice and image features to Plus and Enterprise users over the next two weeks, allowing time to refine safety measures and prepare users for more advanced AI.

OpenAI adds: “Our goal is to build safe and beneficial AGI. We believe incremental tool releases let us improve and mitigate risks over time, while preparing everyone for future, more powerful AI systems.”

Conclusion

Nearly a year after its launch, ChatGPT continues to gain new features while striving to avoid new problems, offering fresh solutions.

As more people adopt voice control and image search, and ChatGPT moves closer to becoming a truly multimodal, useful virtual assistant, the technical barriers to entry will keep rising.