ChatGPT 4 is coming as early as next week and will likely go with a new and potentially dreadful feature: video.
Currently, ChatGPT and Microsoft’s updated Bing search engine are powered by ChatGPT 3.5 large language models, which allows them to respond to questions in a human-like way. But both AI implementations have had their fair share of problems so far, so what can we expect, or at least hope to see, with a new version on the horizon?
According to Microsoft Germany’s CTO, Andreas Braun (as reported by Neowin (opens in new tab)), the company “will introduce GPT 4 next week, where we will have multimodal models that will offer completely different possibilities – for example, videos.” Braun made the comments during an event titled ‘AI in Focus – Digital Kickoff’.
Essentially, AI is definitely not going away anytime soon. In its current state, we can interact with OpenAI’s chatbot strictly through text, providing inputs and controls and getting conversational, mostly helpful, answers.
So the idea of having ChatGPT-powered chatbots, like the one in Bing, being able to reply in other mediums other than plain text is certainly exciting – but it also fills me with a bit of dread.
As I mentioned earlier, ChatGPT’s early days were marked with some strange and controversial responses that the chatbots gave to users. The one in Bing, for example, not only gave out incorrect information, but it then argued with the user who pointed out its mistakes, causing Microsoft to hastily intervene and limit the amount of responses it can provide in a single chat (and which Microsoft is only now slowly increasing again).
If we start seeing a similar streak of weirdness with videos, there could be even more concerning repercussions.
Ethics of AI
In a world where AI-generated ‘deepfake’ videos are an increasing concern for many people, especially those who unwittingly find themselves starring in those movies, the idea of ChatGPT dipping its toes into video creation is a bit worrying.
If people could ask ChatGPT to create a video starring a famous person, that celebrity would likely feel violated. While I’m sure many companies using ChatGPT 4, such as Microsoft, will try to limit or ban pornographic or violent requests, the fact that the ChatGPT code is easily available could mean more unscrupulous users could still abuse it.
There’s also the matter of copyright infringement. AI generated art has come under close scrutiny over where it is taking its samples from, and this will likely be the case with videos as well. Content creators, directors and streamers will likely take a dim view of their works being used in AI generated videos, especially if those videos are controversial or harmful.
AI, especially ChatGPT, which only launched a few months ago, is still in its infancy, and while its potential has yet to be fully realised, so too have the moral implications of what it can achieve. So, while Microsoft’s boasts about video coming soon to ChatGPT is impressive and exciting, the company also needs to be careful and make sure both users and original content creators are looked after.