Yesterday was Andreas Braun, CTO of Microsoft Germany quoted saying that GPT-4 will be introduced next week and will include multi-modal models. The report, which appeared in German news outlet Heise, immediately sparked renewed online chats about the possibility of GPT-4’s debut less than four months after the release of the GPT 3.5 series, which ChatGPT is tuned to.
Coincidentally, deep learning pioneer Yoshua Bengio, who developed the Turing Prize 2018 also made comments yesterday with Geoffrey Hinton and Yann LeCun about ChatGPT and the potential of multimodal models.
In a virtual Q&A titled “What is Missing in ChatGPT? Bridging the gap to human-level intelligence,” Bengio said current work on multimodal large neural networks, which have images or video as well as text, “would help a lot” with the “world model” problem — that is, that models need to understand the physics of our world.
He also warned that market pressure is likely to force tech companies to secrecy rather than openness with their AI models, and that the “media frenzy” surrounding ChatGPT is a “wake-up call” about the potential of powerful AI systems to both do good for society and raise important ethical concerns.
ChatGPT has raised awareness of the potential of powerful AI
Bengio emphasized that while impressive, ChatGPT is a “very small step” scientifically, calling it “primarily a technical advancement.” ChatGPT is more important from a social point of view, he explained, i.e. making people aware of what can be done with AI.
But, he warned, it’s up to people to decide how to design these machines – which, to some extent, Turing test — from an ethical and responsible point of view.
“Are we going to build systems that are going to help us live better lives in a philosophical sense, or will it just be a tool of power and profit?” he said.
The need for regulation
In our economic and political system, “the right answer to this is regulation,” he said, pointing out that startups are willing to take risks that lead to bigger Big Tech companies like Google and Microsoft feeling “forced to jump into the race.” “.
Protecting the public, he added, “is good for everyone in the long run and it levels the playing field — so that the companies more willing to take risks with the public’s welfare are not rewarded for it.”
He stressed that discussions are underway to ensure that AI regulation does not harm the innovation economy. “But it’s going to slow down some things, but that’s probably a good thing,” he said.
A long-term view on ChatGPT and LLMs
Bengio recognized that companies currently feel the urgency to incorporate ChatGPT and other LLMs into their products and services. But he pointed out that academics and some companies are also looking to the future.
“How do we become the next big company in the field? How do we lead? For that you have to think about what is missing, what the failure modes are,” he said. “That kind of research is difficult and can take years to answer. Hopefully some people will have the vision to see beyond the immediate panic that I think is going on right now.”
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.