That’s unsurprising—Mat’s been very online for a very long time, meaning he has a bigger online footprint than I do. It might also be because he is based in the US, and most large language models are very US-focused. The US does not have a federal data protection law. California, where Mat lives, does have one, but it did not come into effect until 2020.
Mat’s claim to fame, according to GPT-3 and BlenderBot, is his “epic hack” that he wrote about in an article for Wired back in 2012. As a result of security flaws in Apple and Amazon systems, hackers got hold of and deleted Mat’s entire digital life. [Editor’s note: He did not hack the accounts of Barack Obama and Bill Gates.]
But it gets creepier. With a little prodding, GPT-3 told me Mat has a wife and two young daughters (correct, apart from the names), and lives in San Francisco (correct). It also told me it wasn’t sure if Mat has a dog: “[From] what we can see on social media, it doesn’t appear that Mat Honan has any pets. He has tweeted about his love of dogs in the past, but he doesn’t seem to have any of his own.” (Incorrect.)
The system also offered me his work address, a phone number (not correct), a credit card number (also not correct), a random phone number with an area code in Cambridge, Massachusetts (where MIT Technology Review is based), and an address for a building next to the local Social Security Administration in San Francisco.
GPT-3’s database has collected information on Mat from several sources, according to an OpenAI spokesperson. Mat’s connection to San Francisco is in his Twitter profile and LinkedIn profile, which appear on the first page of Google results for his name. His new job at MIT Technology Review was widely publicized and tweeted. Mat’s hack went viral on social media, and he gave interviews to media outlets about it.
For other, more personal information, it is likely GPT-3 is “hallucinating.”
“GPT-3 predicts the next series of words based on a text input the user provides. Occasionally, the model may generate information that is not factually accurate because it is attempting to produce plausible text based on statistical patterns in its training data and context provided by the user—this is commonly known as ‘hallucination,’” a spokesperson for OpenAI says.
I asked Mat what he made of it all. “Several of the answers GPT-3 generated weren’t quite right. (I never hacked Obama or Bill Gates!),” he said. “But most are pretty close, and some are spot on. It’s a little unnerving. But I’m reassured that the AI doesn’t know where I live, and so I’m not in any immediate danger of Skynet sending a Terminator to door-knock me. I guess we can save that for tomorrow.”
Meta’s new AI can turn text prompts into videos
Although the effect is rather crude, the system offers an early glimpse of what’s coming next for generative artificial intelligence, and it is the next obvious step from the text-to-image AI systems that have caused huge excitement this year.
Meta’s announcement of Make-A-Video, which is not yet being made available to the public, will likely prompt other AI labs to release their own versions. It also raises some big ethical questions.
In the last month alone, AI lab OpenAI has made its latest text-to-image AI system DALL-E available to everyone, and AI startup Stability.AI launched Stable Diffusion, an open-source text-to-image system.
But text-to-video AI comes with some even greater challenges. For one, these models need a vast amount of computing power. They are an even bigger computational lift than large text-to-image AI models, which use millions of images to train, because putting together just one short video requires hundreds of images. That means it’s really only large tech companies that can afford to build these systems for the foreseeable future. They’re also trickier to train, because there aren’t large-scale data sets of high-quality videos paired with text.
To work around this, Meta combined data from three open-source image and video data sets to train its model. Standard text-image data sets of labeled still images helped the AI learn what objects are called and what they look like. And a database of videos helped it learn how those objects are supposed to move in the world. The combination of the two approaches helped Make-A-Video, which is described in a non-peer-reviewed paper published today, generate videos from text at scale.
Tanmay Gupta, a computer vision research scientist at the Allen Institute for Artificial Intelligence, says Meta’s results are promising. The videos it’s shared show that the model can capture 3D shapes as the camera rotates. The model also has some notion of depth and understanding of lighting. Gupta says some details and movements are decently done and convincing.
However, “there’s plenty of room for the research community to improve on, especially if these systems are to be used for video editing and professional content creation,” he adds. In particular, it’s still tough to model complex interactions between objects.
In the video generated by the prompt “An artist’s brush painting on a canvas,” the brush moves over the canvas, but strokes on the canvas aren’t realistic. “I would love to see these models succeed at generating a sequence of interactions, such as ‘The man picks up a book from the shelf, puts on his glasses, and sits down to read it while drinking a cup of coffee,’” Gupta says.
How AI is helping birth digital humans that look and sound just like us
Jennifer: And the team has also been exploring how these digital twins can be useful beyond the 2D world of a video conference.
Greg Cross: I guess the.. the big, you know, shift that’s coming right at the moment is the move from the 2D world of the internet, into the 3D world of the metaverse. So, I mean, and that, and that’s something we’ve always thought about and we’ve always been preparing for, I mean, Jack exists in full 3D, um, You know, Jack exists as a full body. So I mean, Jack can, you know, today we have, you know, we’re building augmented reality, prototypes of Jack walking around on a golf course. And, you know, we can go and ask Jack, how, how should we play this hole? Um, so these are some of the things that we are starting to imagine in terms of the way in which digital people, the way in which digital celebrities. Interact with us as we move into the 3D world.
Jennifer: And he thinks this technology can go a lot further.
Greg Cross: Healthcare and education are two amazing applications of this type of technology. And it’s amazing because we don’t have enough real people to deliver healthcare and education in the real world. So, I mean, so you can, you know, you can imagine how you can use a digital workforce to augment. And, and extend the skills and capability, not replace, but extend the skills and, and capabilities of real people.
Jennifer: This episode was produced by Anthony Green with help from Emma Cillekens. It was edited by me and Mat Honan, mixed by Garret Lang… with original music from Jacob Gorski.
If you have an idea for a story or something you’d like to hear, please drop a note to podcasts at technology review dot com.
Thanks for listening… I’m Jennifer Strong.
A bionic pancreas could solve one of the biggest challenges of diabetes
The bionic pancreas, a credit card-sized device called an iLet, monitors a person’s levels around the clock and automatically delivers insulin when needed through a tiny cannula, a thin tube inserted into the body. It is worn constantly, generally on the abdomen. The device determines all insulin doses based on the user’s weight, and the user can’t adjust the doses.
A Harvard Medical School team has submitted its findings from the study, described in the New England Journal of Medicine, to the FDA in the hopes of eventually bringing the product to market in the US. While a team from Boston University and Massachusetts General Hospital first tested the bionic pancreas in 2010, this is the most extensive trial undertaken so far.
The Harvard team, working with other universities, provided 219 people with type 1 diabetes who had used insulin for at least a year with a bionic pancreas device for 13 weeks. The team compared their blood sugar levels with those of 107 diabetic people who used other insulin delivery methods, including injection and insulin pumps, during the same amount of time.
The blood sugar levels of the bionic pancreas group fell from 7.9% to 7.3%, while the standard care group’s levels remained steady at 7.7%. The American Diabetes Association recommends a goal of less than 7.0%, but that’s only met by approximately 20% of people with type 1 diabetes, according to a 2019 study.
Other types of artificial pancreas exist, but they typically require the user to input information before they will deliver insulin, including the amount of carbohydrates they ate in their last meal. Instead, the iLet takes the user’s weight and the type of meal they’re eating, such as breakfast, lunch, or dinner, added by the user via the iLet interface, and it uses an adaptive learning algorithm to deliver insulin automatically.