My AI Predictions for 2024

As 2023 draws to a close, I’m going to make my predictions for 2024. Let’s see how close I get.

RESPONSIBLE AI

12/29/20232 min read

Robot gazing into a crystal ball
Robot gazing into a crystal ball

2023 was a breakout year for AI. Google’s Bard generative AI estimates that the phrase “AI” is used in over 1 million news articles in 2023, making it the most frequently used term in news media. I also asked Bard to make some predictions for 2024, and it basically wrote that 2024 would be like 2023 except more so. Which is perhaps the problem with GenAI: it only tells us what we want to hear without any thought of what’s it’s actually saying.

On the other hand, I put thought into my predictions, but I’m likely to be equally right as Bard. This is essentially the problem with humans: they think they’re right more often than they actually are right. Here are my predictions.

  • AI equity will not be shared further than US, EU and China. Others will continue to be left behind.

  • Open AI’s market capitalisation will exceed that of its biggest investor, Microsoft. (As of December 2023, Open AI’s market cap is ~$80B whereas Microsoft’s is $2.3T.)

  • There will be a single event of catastrophic destruction of trust. This will cause open source foundation model repos like Hugging Face to be driven out of business by legal jeopardy. But this won’t stop Open AI’s rise because big tech will face down competition regulators.

  • At the time of the US election, there will be more synthetic content than actual content created by creators published. Not all will be disinformation, most will be advertising, a little will be actual programming, but crucially we won’t be able to tell the difference between the three.

  • There will be a first prosecution under the UK’s online safety bill, and it will fail. There will be actual harm cause by GenAI yet no ability to redress.

  • Cost of training a next generation foundation model will exceed $1B. Which will drive concentration of AI power through consolidation of models with chips. This will cause geopolitical jeopardy like China eying Taiwan hungrily during its election.

  • Smaller companies will not be able to afford to own their full AI supply chain and “do AI properly”. But being trapped in this constant hype cycle of AI being the most powerful tool every invented, they will still put investment into finding a problem for it to solve. They will either buy from a new breed of service providers sourcing data and pre-trained models - or just use Chat GPT.

This is all a bit gloomy, so here’s a bonus prediction that’s a little more hopeful.

  • There will be more citizen voices around the world calling for ethical use of AI. Governments will begin to listen, realise that ethical AI aligns with their need for jurisdictional governance and digital sovereignty, and start to implement AI legislation. The first of these will be the EU AI Act 2024.

Wishing you health, happiness and good fortune in 2024!