Neural Notes: OpenAI drama, Musk launches Grok (not that one), Aussie watchdogs on AI

Welcome back to Neural Notes: A weekly column where I condense some of the most interesting and important AI news of the week. In the latest edition, we have the OpenAI drama from the past week and Elon Musk taking the opportunity to launch his own rival to ChatGPT. Meanwhile, here in Australia, the Digital Platform Regulators Forum (DP-REG) has fired up some working papers and a website to address the regulation of digital platform technologies.

How about that OpenAI drama?

We can’t talk about AI this week without getting into the popcorn-munching saga that has been going down in the valley. Over the past week, the troubled company has seen its board fire CEO Sam Altman, had 500 employees threaten to revolt, seen the Microsoft CEO weigh in, Salesforce weigh in — all of which ultimately resulted in Altman being hired back with a brand new (and all male, just sayin’) board.

There was a LOT more that went down and luckily for you, we already published a timeline for you to peruse.

But even if you followed the story, I recommend taking a look anyway because I also took a deep dive into the unique board structure and corporate governance over at OpenAI that made it so easy for Altman to be fired in the first place. It’s incredibly interesting.

Elon Musk enters the chat with Grok

Earlier this year Elon Musk launched his own AI company, xAI, to ‘help understand the universe’. No, we don’t know what that meant either.

Musk, who co-founded OpenAI, has been openly critical of the direction it went in. And his more philosophical approach to his own company seemed to indicate wanting to take AI in a different direction.

However, rumours that he was spinning up a rival to ChatGPT seem to be at least a little accurate. In a bit of convenient timing, Musk confirmed this week that he will be launching his recently announced chatbot, Grok, which will be integrated into the X web application.

The name is of course quite funny for us Aussies who associate it with Mike Cannon-Brooke’s venture firm.

https://twitter.com/elonmusk/status/1727141345876402236

Grok is set to be available to X Premium+ subscribers from next week. The release follows a sneak peek revealed by app researcher Nima Owji earlier in the week.

According to the xAI site, Grok isn’t just a rival to the likes of ChatGPT and Bard — it’s positioning itself as a superior offering.

“Grok is an AI modeled after The Hitchhiker’s Guide to the Galaxy, so intended to answer almost anything and, far harder, even suggest what questions to ask,” the xAI website reads.

It also warns potential users that Grok sure is a product from an Elon Musk company: “Grok is designed to answer questions with a bit of wit and has a rebellious streak, so please don’t use it if you hate humor!

“It will also answer spicy questions that are rejected by most other AI systems.”

xAI also claims that Grok has real-time knowledge of the world “thanks to the X platform”. If true, this is a direct comparison to the likes of ChatGPT which is a couple of years behind (with the exception of GPT-4 Turbo, which closes the knowledge gap to earlier in 2023).

Unsurprisingly, Grok’s integration promises to bolster X’s Premium subscription service, which has been underperforming. X recently diversified its subscription options, introducing a three-tier system with varying benefits and prices.

The top tier, Premium+, will include Grok, potentially driving up subscriptions amid X’s need for increased revenue due to a recent advertiser retreat.

Aussie watchdogs weigh in on AI

Bringing things back to Australia, the Digital Platform Regulators Forum (DP-REG) has released working papers on algorithms and large language models (LLMs) in generative AI. These papers coincide with the launch of a dedicated DP-REG website.

DP-REG — which is a collaboration between the ACCC, ACMA, eSafety Commissioner, and OAIC — aims to streamline the regulation of digital platform technologies in Australia. The papers align with DP-REG’s 2023/24 strategic priorities to understand the impact of algorithms and assess generative AI’s benefits and risks.

The first paper provides a literature summary on the harms and risks of algorithms, enhancing members’ understanding of algorithm types relevant to their regulatory roles.

The second paper dives into LLMs and their potential regulatory implications. These include the need to balance innovation as well as safety in the space, jurisdictional challenges and simply trying to keep up with the rapid technology.

I still need to trawl through the entirety of both papers, but I’ll report back if I find anything particularly interesting or eyebrow-raising. In the meantime, it’s good to see something being published by the who’s who of Aussie regulators.

Regulation has been one of the biggest concerns regarding the AI renaissance happening over the past year. As we all know, technology always moves faster than the law — and that seems all the more true when it comes to AI. We have already seen the implications of generative AI in particular being the Wild West when it comes to content creation.

While we have seen some incredible innovations from the likes of Canva, Adobe and Salesforce — we have equally seen issues arise when it comes to the works of artists and authors being used without permission or outright being stolen.

And then there’s the added complication of businesses needing to know that assets they create with the use of AI will be both free of copyright materials and also protected as their own IP.

As a very boring nerd, I am fascinated to see how this unfolds further, especially with the watchdogs getting more involved.