OpenAI to Update ChatGPT: Supporting Image and Voice Input

Google's top executives may feel a bit uneasy with the emergence of ChatGPT by the end of 2022. This emerging chatbot answers questions in an ad-free and pleasant manner, posing a clear threat to Google's search business. However, due to its limited access to the internet and Google's control over the industry, it remains just a fun toy for most people. Now, ChatGPT operated by OpenAI challenges Google once again by integrating voice and image queries into its popular chatbot.

According to reports, OpenAI has announced the upcoming release of a new version of ChatGPT that allows users to ask questions to the chatbot through voice. The service converts voice into text, passes it to ChatGPT, and then converts the answers into audio to be read to the user. The company states that users can choose from five "human-like audio" options. More audio options may be provided in the future, but currently, this is just the starting point for the company in terms of voice-to-text conversion.

Additionally, users can also upload photos to ask questions. Although this may seem risky, ChatGPT may already have some built-in detection capabilities. This raises questions about what comments ChatGPT will make about people in the photos. OpenAI states that they have limited ChatGPT's ability to comment on people, but there are clearly some gray areas. For example, users can upload a photo of a former president and ask about their legacy, or upload a blurry photo of a neighbor.

OpenAI states that they developed this feature to allow users to upload photos and ask questions to help improve the answers provided. In the example mentioned above, a user uploaded a photo of a bicycle and asked ChatGPT how to lower the seat height. The chatbot would provide specific guidance and then ask for another photo of the seat height adjustment mechanism for further guidance. Currently, Google Lens also offers similar services but lacks the ability to further optimize queries.

OpenAI states that these features will be initially launched for paid users and will be available to free users two weeks later. However, these features also raise concerns about privacy, accuracy, and whether the chatbot will occasionally "hallucinate" like other similar services.