Latest ChatGPT update makes it smarter, but also a little creepier
AI used for photos isn’t exactly new. Many companies, such as Apple and Google, already use AI to a certain extent when it comes to managing photos on our phone. It is used to identify objects and people, for example, which makes sorting and searching for images a lot easier.
But it looks like OpenAI’s ChatGPT will be able to do a much better, and smarter, job at that. OpenAI has announced its latest version of its AI, GPT-4, which will now be able to look at a photo and analyze it beyond just pointing out what’s in the photo. For example, it can look at a photo and when asked to describe what’s strange about it, it will be able to point out the differences.
This is actually kind of mind-blowing when you think about it because first of all, the AI needs to know what’s in the photo and understand the context. Secondly, it also needs to understand what makes an image “weird” or “strange” in terms of human context. In an example image they used, it showed a man hanging from the back of a taxi ironing a shirt in the middle of the street, which we imagine that most people would find strange.
It can even understand jokes and memes and explain it, which once again is very impressive because it means that the AI can now think like a human and understand what would make an image funny or interesting. This means that in the future, sorting and searching for images could get a lot smarter and easier.
For example, you might be able to search your library for things like, “Show me an image of myself at a party looking awkward” and it might be able to give you some useful results. But what do you think? Useful, or a little creepy?
Source: OpenAI