View all on-demand sessions from the Intelligent Security Summit here.
I’m going south tomorrow morning. Straight across I-95, from central New Jersey to northeastern Florida, where I’ll be setting up my laptop in St. Augustine for the next two months. It’s about as far from Silicon Valley as I can be in the continental US, but there you’ll see me gearing up for the first artificial intelligence (AI) news of 2023.
Here are the 5 biggest AI stories I’m waiting for:
1. GPT-4
ChatGPT is So 2022, don’t you think? The Hype Around OpenAI’s Chatbot “research example”, released on November 30, has barely reached its peak, but the boisterous speculation of what comes next – GPT-4 – is like the sound of millions of Swifties waiting for Taylor’s next album.
If expert predictions and OpenAI’s cryptic tweets are correct, GPT-4 – with more parameters and trained on more data – will debut in early to mid-2023 and “ghosts will be blown up.” It will still be filled with the unreliable “plausible BS” of ChatGPT and GPT-3, but it will possibly be multimodal – able to work with images, text and other data.
It’s not even three years ago GPT-3 was released, and only two since the publication of the first DALL-E research article. When it comes to the rate of innovation for major language models in 2023, many say “brace yourself.”
2. The EU AI law
AI technology is advancing rapidly, but so is AI regulation. While there have been several state-based AI-related bills passed in the US, it is bigger government regulation – in the form of the EU AI law – that everyone has been waiting for. On December 6, the EU AI law progressed a step towards law, after which the Council of the EU passed its amendments to the draft law, opening the door for the European Parliament to “finalize their common position.”
According to Avi Gesser, partner at Debevoise & Plimpton and co-chair of the firm’s Cybersecurity, Privacy and Artificial Intelligence Practice Group, the EU AI law seeks to create a risk-based regime to address the riskiest outcomes of artificial intelligence . intelligence. As with the GDPR, it will be an example of a comprehensive European law taking effect and slowly seeping into several state and sector-specific laws in the US, he recently told VentureBeat.
Boston Consulting Group calls the EU’s AI law “one of the first comprehensive regulatory frameworks for AI” and expects it to become law in 2023. almost every company.
3. The battle of search
Last week, the New York Times called ChatGPT a “code red” for Google’s search activity. And in mid-December, You.com announced that it had opened up its search platform to generative AI apps. Then, on Christmas Eve, You.com debuted YouChatwhich it called “Conversational AI with citations and real-time data right in your search bar.”
To me, this all boils down to what could be a real battle for the future of search in 2023 – I’m already munching on popcorn waiting for Google’s next move. As I recently wrote, Google processes billions of searches every day, so it won’t be going anywhere for a while. But perhaps ChatGPT — and even You.com — is just the beginning of a new, imaginative way of thinking about the future of AI and search.
And as Alex Kantrowitz said Axios recently, Google may need to take a step: “It’s game time for Google,” he said. “I don’t think it can sit on the sidelines for too long.”
4. Open source versus closed AI
I am fascinated by the ongoing discussion around open source and closed AI. With the rise of Hugging Face’s open source model development, the company reached a valuation of $2 billion in May; Stable diffusions big summer splash in the text-to-image space; and the first open source copyright lawsuit focused on GitHub CoPilot, open source AI had a big, influential year in 2022.
That will certainly continue into 2023, but I’m particularly interested in how it compares to the evolution of closed source AI models. Finally, OpenAI shifted to closed source and is now about to release GPT-4, arguably the most highly anticipated AI model ever – which is certainly a competitive edge, right?
On the other hand, MIT Technology Review predicts “An open source revolution has begun that matches, and sometimes exceeds, what the richest labs are doing.” Sasha Luccioni, research scientist at Hugging Face, agreed, adding that open source AI is more ethical. She tweeted last week that open sourcing AI models “make it easier to find and analyze ethical issues, instead of keeping them closed source and saying ‘trust me, we filter out all the bad stuff’.
5. Is AI running out of training data and computing power?
Will 2023 herald an AI era of creative conservation when it comes to data and compute?
ChatGPT’s computational cost is “eye-popping” according to OpenAI Sam Altman, while says IBM that we’ve run out of computing power altogether, because while AI models are “growing exponentially,” the hardware to train and run them hasn’t progressed that fast. Meanwhile, a research paper progress that “data typically used for language model training may be used up in the near future — as early as 2026.”
I’m curious how this will play out in the coming year. In the end, won’t big equal better when it comes to data and compute? Will new AI chips designed for deep learning models change the game? Will synthetic data be the answer to the training problem? I have my popcorn ready for this one too.
I wish you all a happy new year!
On January 2nd I will be back in my temporary ‘office’ on the beach. Until then, enjoy the last week of 2022 and a happy, healthy new year. As a reminder, I’m on Twitter at @sharongoldman and can be reached at [email protected].
VentureBeat’s mission is to become a digital city plaza where tech decision makers can learn about transformative business technology and execute transactions. Discover our Briefings.
