We asked ChatGPT what the new or improved features of version -4o are. Here is the answer:
ChatGPT: Summary of new or improved features in ChatGPT-4o:
- language functions:
- read-aloud functionChatGPT can read texts aloud, which helps you obtain information without having to read.
- voice inputYou can enter questions and commands directly by voice, without having to type.
- Better data analysis:
- ChatGPT can now analyse files better, even multiple files at once, to give you helpful insights.
- Customisable responses:
- You can give ChatGPT personal instructions so that the answers better suit your wishes and needs.
- Increased security:
- Enhanced security measures ensure that ChatGPT responds in a safer and more responsible manner and avoids harmful content.
- user-friendliness:
- New apps for Android and iOS make it easier to use ChatGPT on your mobile phone or tablet..
__________________
Much of this can be summarised by saying that ChatGPT-4o's responses are simply better, more precise and more accurate. However, its ability to communicate using language is outstanding. We tried it out in the editorial office. It wasn't that easy, but in the end we managed it, even without outside help. Well, that's not entirely true. We asked ChatGPT how voice communication between us works and what settings we need to configure. That's how we managed it. Here is a brief, hopefully generally understandable explanation of what to do. In the editorial office, we use Apple i-Macs and the Firefox and Safari browsers.
- First, you have to teach your computer that inputs can now also be made by voice and not just as typed text. This can be done in the system settings under „Keyboard“. There you will find the settings that allow dictation. This must be activated.
- Then the microphone must be activated. It can be an integrated microphone (as with the i-Mac) or an external one.
- Once this has been done, you can, in principle, make entries using your voice. Click in the usual ChatGPT input field to activate it. In our case, double-click on the key labelled „fn“. A microphone symbol will appear on the screen. You can now speak your input. When you are finished, double-click the fn key to end voice input. Now you must submit the text (perhaps after proofreading it) as usual by clicking the Enter key. In Windows, instead of fn, it is the Windows key + H; Chrome uses Google's dictation feature.
- ChatGPT has now received your request and will respond shortly. This usually takes one to three seconds. The response will initially be in text form. At the end of the response, ChatGPT will send a line with grey symbols. On the far left of this line is a speaker icon. Click on it and the response will be read aloud to you. It is not yet possible to choose between different voices, but ChatGPT uses a very pleasant female voice that is easy to understand and has an almost human intonation.
- Of course, this is not yet a real-time conversation, because you always have to double-click on fn twice for your questions – also so that the microphone does not remain permanently activated – and you have to start listening by clicking on the speaker icon. But with a little practice, the clicks become routine. Yes, phonetic questions and answers can definitely be described as a conversation.
Please do not send us any questions about how this works on the device you are using: PC, tablet or mobile phone. Windows, macOS, Chrome, Linux, Android. Firefox, Chrome, Safari. Everyone has a different system for going online. We cannot provide technical assistance on how to set up a voice conversation with ChatGPT. But we don't have to. Just ask the computer specialist ChatGPT! Enter which device you are using, which operating system, which browser – if possible, with the version number. ChatGPT will immediately tell you what settings you need to make in order to talk to ChatGPT. It worked for us – you can do it too. Provided you have a microphone and your speaker activated. We wish you stimulating conversations with ChatGPT, the nice „artificial“ lady. And treat her well! Friendly questions get friendly answers. Just like in real life.
Finally, we would like to answer the question in the headline of this series of articles. The day ChatGPT learned to speak was 13 May 2024.