As we embrace the new year (I won't miss 2023!), I've been thinking a lot about artificial intelligence and what 2024 holds. Last year took a lot of people by surprise with a whirlwind of advancements and hype, but as with all technology, there comes a moment of equilibrium.

Here are my key predictions for AI in 2024:

A Plateau in Progress

Remember how we'd all be riding in self-driving cars by now?

In 1995, Carnegie Mellon researchers took their self-driving car, called NavLab 5, to the road, traveling 2,797 miles from Pittsburgh to San Diego.

We still don't have fully autonomous vehicles. This is because the final stages of any product are the most challenging, and AI is no exception. While 2023 was a year of significant advancements, we should anticipate incremental improvements in 2024 unless an unexpected breakthrough occurs.

Again, I'll remind my audience that I don't worry about AI replacing developer jobs because if they get that good, all jobs are replaceable, and we will have a new social contract.

Challenges in Politics and Cybersecurity

One of the more pressing concerns is AI's impact on politics and cybersecurity. The ease of generating phishing attacks and spreading misinformation will reach new heights, exacerbated by state actors employing LLMs for mass internet spamming.

This trend underscores the need for vigilance, robust countermeasures, and trying our best to teach critical thinking in our schools.

With 2024 being a presidential election year in the U.S., I'm predicting huge efforts and funds spent on spreading misinformation through platforms like X, Facebook, and Reddit.

The Rise of "Verified Human" Content

As AI becomes more pervasive, there will be a growing appreciation for genuine human interactions and creations. Platforms like YouTube are already implementing measures to label AI-generated content.

This shift will prompt consumers to seek out what I term "verified human" content – be it videos, blog posts, or product reviews – valuing authenticity and the human touch.

The Oversaturation (and Fall) of AI Startups

Many startups, primarily those offering simple extensions of existing LLMs like ChatGPT, face the risk of failure. These products are easily replicable, may be internalized by larger providers, and the technology still requires significant human oversight, so I'm unconvinced it can deliver on the hype.

This doesn't mean that companies shouldn't be evaluating opportunities and product enhancements. Still, I doubt that first-mover advantage will be significantly valuable based on the types of ventures I see.

Overall, I'm very curious how 2024 will shake out, especially to see how regulators deal with (or don't) the nefarious uses that LLMs can be put to at scale.