Where Is AI Going Next?

For the last few weeks, it’s clear that there are many AI models in training for release in this year, including ChatGPT-5, likely to be the most powerful of them all. Now that we are firmly into 2024, I thought it would be interesting to think about three advancements we are likely to see this year, particularly against the backdrop of geopolitical upheaval across the world and the role that AI might play in that, for good or ill. Lets dive in…

Let me start off by saying that I think AI improvements this year will be tempered somewhat by one single thing: the US Presidential election. I would imagine that none of the big tech companies driving the adoption of AI will want any of their more advanced tools to be exploited during what will undoubtedly be the most chaotic election in history. They will not want to catch any flack from that so I think the really big improvements will happen at the end of the year. Nonetheless, that doesn’t mean that nothing is happening, rather that training and safety testing will be even more deliberate and thorough. I imagine ChatGPT-5 will be released late November as it has been confirmed that it is in training right now. That said, AI is still accelerating at a pace that's nothing short of breathtaking and is unfolding in spectacular new dimensions, from text to video to enhanced multimodal interactions and groundbreaking advancements in AI driven robotics.

Text to video is set to become a cornerstone of content creation. This technology is not just a new tool in the artist's palette; it's an entirely new canvas. Companies like RunwayML, InVideo and many others are revolutionising the film and advertising industries to creating educational content that can visually depict complex concepts. It's like having a film studio at your fingertips, where your words are the script and AI is the director. Expect making a movie to go from incredibly complex and difficult to relatively simple in the nearish future.

The AI of 2024 will speak our language, literally and figuratively. Speech in, speech out. Gone will be the days of clunky text inputs and text or robotic voice outputs. This is about creating a more natural, intuitive way to interact with technology. Imagine conversational AI tutors that can understand a student's frustration through their tone of voice and respond with encouraging, personalised guidance. Companies like Eleven Labs are at the vanguard of this technology with others joining the fray daily. Of course, deepfake audio will also be something that we will have to grapple with all year.

Here’s an example of where this technology is today:

Robotics in 2024 isn't just about more sophisticated single task machine like we have seen for several years in automotive manufacture. Instead it will be about humanoid robots that think, learn, and adapt, perfecting skills that I have never been able to master!

However, the current pinnacle of the fusion betrween AI and robotics is the most unlikely robot of all, the car.

That’s it for another Saturday. I hope you found the thoughts and videos interesting. As always, keen to hear your comments at [email protected]