The shifting AI realm: What’s next on the horizon

Governments didn’t really make much headway in regulating artificial intelligence this year, even though the issue of its safety became a hot topic globally, as per the 2023 “State of AI” report released recently. The report mentioned that AI safety finally stepped into the limelight in 2023, leaving behind its previous image as a neglected relative in the world of AI research.

However, with no clear agreement worldwide on how to move ahead with regulation, the creators of advanced AI systems were trying to influence the rules themselves by suggesting their own ways of regulation. The report pointed out that in 2023, Big Tech regained control over the AI sector, even though open-source initiatives seemed to be leading the way last year.

Also Read: Can you bypass ChatGPT restrictions and get an unfiltered AI chatbot?

Strong computing has amplified AI models

This year, with a persistent lack of powerful computer chips, the big tech companies had the upper hand, thanks to their strong computing setup and ample financial resources, as the cost of training massive AI models kept going up.

Nathan Benaich, the report’s author, informed TIME that last year, many individuals were coming together on Discord servers, plenty of open-source models were emerging, and it appeared that the Big Tech companies weren’t particularly active.

Nathan Benaich noted that this year, there was a noticeable swing in the opposite direction, as almost every public tech company was working on creating or incorporating AI systems into their products. The open-source community remained active and was swiftly trying to match the capabilities of closed-source technologies, but it wasn’t immediately clear how they could completely replicate GPT-4.

“Last year we saw a lot of people assembling in Discord servers, we saw a lot of open source models, and it didn’t seem like Big Tech companies were doing all that much,” Nathan Benaich said. “This year, it looks like a pretty significant snap back in the other direction, with pretty much every public tech company making moves to develop or integrate AI systems into their products. The open-source world is still very vibrant and is rapidly trying to catch up with closed-source capabilities, but it doesn’t look immediately obvious how you’d 100% clone GPT-4.”

Open AI’s GPT-4 is leading the way

Now in its sixth year, the State of AI report has become a go-to guide for the AI industry, highlighting trends and making forecasts for the upcoming year. This year, Benaich, an investor at Air Street Capital, put the report together. In the past, Ian Hogarth, an investor now leading the U.K. government’s AI safety taskforce, co-authored it.

According to the report, OpenAI’s GPT-4 still holds the top spot as the most powerful large language model (LLM) eight months after its launch, outperforming every other LLM on traditional benchmarks and tests crafted to assess human capabilities.

However, the report highlights that it’s getting trickier to compare state-of-the-art AI systems as they become more robust and adaptable. The industry is increasingly leaning toward a “vibes-based” method to assess LLMs, as the traditional tests, known as benchmarks, are becoming less conclusive, according to the report.

Also Read: EU’s AI Act might be bad news for small firms 

Nobody wants to share AI knowledge

The report mentions that in 2023, the trend of AI companies openly sharing their cutting-edge research came to a halt. According to the report, OpenAI refused to share any substantial details about the system architecture of GPT-4, and both Google and Anthropic made similar choices regarding their models.

The report suggests that with increasing economic stakes and safety concerns (you can believe what you want), previously transparent companies are now adopting a more secretive culture concerning their most advanced research.

“As the economic stakes and the safety concerns are getting higher (you can choose what to believe), traditionally open companies have embraced a culture of opacity about their most cutting-edge research,” the report says.