The AI Daily Brief: Artificial Intelligence News and Analysis - Can Viral NSFW Taylor Swift AI Images Make Deepfake Issues Go Mainstream?

Episode Date: January 25, 2024

If anyone can make AI deepfakes a mainstream issue it's Taylor Swift. Plus geopolitical competition around compute access heats up. ** LEARN AI THIS YEAR! Registration is very briefly open for our February cohort of our AI Education Beta program. Get access to a library of 60+ tutorials, case studies and challenges New lessons drop every week day Join a passionate community of like-minded learners Topics include LLMs, AI nocode tools, image generators, voice synthesizers, AI for professional applications like presentation generation, website building and more. Learn more and sign up here: https://bit.ly/aibeta Registration closes on Sunday January 28th 11:59pm EST ** ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

Today on the AI breakdown, we're looking at some interesting developments in the geopolitical competition around computing power. Before that on the brief, could Taylor Swift be the reason that AI deepfakes make it to mainstream consciousness? The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.netnetwork for more information about our YouTube, our newsletter, and our Discord. So far, when it comes to the national and global discourse around deepfakes and AI-generated

images of real-world people, much of the discussion has been either, A, theoretical, as in this is a problem that is likely to happen, and so we need to get ahead of it, B, focused on specifically child porn, that's where a huge amount of the policy effort has been, or C, more broadly, has been really focused on elections and politicians. There was, for example, a big dustup in New Hampshire earlier this week when Democratic voters got a robocall from an AI Biden, saying, that they didn't need to go out and vote and that their vote only counted in November. Yet, while there is a lot of coverage of all of that in the media, I would argue that it hasn't really made it to the mainstream discourse. In fact, the AI images that have gotten famous

like the Pope and the Puffer Jacket have seemed more novel, funny, or cool, than harmful and disruptive. Well, that might all be changing right now as a set of very explicit NSFW-Taylor Swift AI images themed around her new relationship. with Travis Kelsey and the NFL went viral yesterday on Twitter. Now, the interesting thing that happened is that when Taylor Swift AI started trending, Swifties took to the platform en masse to post other images of Taylor, real images of Taylor, like from her concerts, with that same text Taylor Swift AI,

as well as the text protect Taylor Swift, to basically push down the offending AI images and make them less easy to access, even though they weren't taken off the platform. I have to say, if you scroll through the results right now, now for Taylor Swift AI, the Swifties have done a pretty remarkable job. Now, of course, this being Taylor Swift and us being in the world that we live in right now, this did make major news. Now, the whole thing was pretty fascinating to me as exemplary of the sort of internet battles that were going

to face a lot in the years to come. Even as Twitter was playing whack-a-mole suspending accounts that had re-shared the images, others were of course popping up, and you can find those images on other sites like Instagram, Reddit, and 4chan as well. Yet the mass army of Swifties who used the medium itself to drive away those images were seemingly as effective as Twitter or X was itself in suspending those accounts that were sharing them. Now, of course, on the one hand, people are angry because this happened to their girl, to Taylor. But at the same time, you have to think that it's going to raise awareness of the broader

issue with this group of people as well. This feels to me like a topic where there may be low-hand. hanging fruit political victories to be had. Many states are exploring additional laws to ban deep fake porn, as well as imposing penalties for those who spread it. This was a topic that was in Biden's executive order last year, and there are multiple pieces of national legislation as well. All I know is if there is one thing that could turn this into a national issue, it is Taylor Swift. Next up, we head over to the world of advertising, where another company has made some big, bold pronouncement about how much it's going to spend, repurposing itself for the AI world.

Publissus, the world's largest advertising conglomerate, has announced that they intend to spend 300 million euros over the next three years, including 100 million euros in this year alone. The goal, they say, is to become, quote, the industry's first AI-powered intelligence system, whatever that means. Now, when it comes to what that money will be spent on, half of the 100 million earmarked for this year will be spent on reskilling and training people inside the company as well as recruiting new talent, while the other half is focused on buying tech, buying software, building out cloud infrastructure and more. I think if you take anything away, the fact that this company is going to spend

$50 million this year re-skilling its employees shows you just how seriously these companies are taking this. Now, speaking of taking things seriously, Google has recently been embroiled in an AI-related chip patent lawsuit, but has finally reached a settlement. Singular computing had brought the lawsuit against Google and was seeking $1.67 billion in damages. Now, Google said that the company had not violated Singular's patent rights, but Singular said that Google had incorporated its technology into processing units that support their AI features across their suite of tools. Basically, Singular's founder Joseph Bates said that he shared his inventions with Google between 2010 and 2014, and that Google's sensor processing units copied those inventions

and infringed on two patents. Now, Google's claim was that the employees who designed at the TPUs never met Bates and created them completely independently. In any case, on the day that closing arguments were scheduled to begin. The companies came to an agreement for a settlement, but that's all the details we have. Meanwhile, the market success of AI continues with IBM becoming the latest company to beat earnings estimates thanks to artificial intelligence. We talked recently about how SAP was restructuring 8,000 jobs in pursuit of more AI competitiveness and how it had beat earnings estimates. And Microsoft also just became the second company to hit a $3 trillion market cap and has over the last year been consistently beating earnings

estimates, especially in its cloud business, thanks to its partnership with OpenAI, and in general, its AI-related sales. In sharing forecasts for full-year revenue growth that were above market estimates, IBM basically said that companies looking to adopt artificial intelligence were bolstering demand for both IBM's IT software as well as its consultancy services. IBM shares were up more than 8% on the news. Finally, Snapchat continues to experiment with different gamified, teen-and-young person focused versions of AI and is apparently going to introduce an AI bitmoji pet. Jonathan Manzano tweeted, Snapchat Plus has added a feature for custom bit moji, allowing you to use AI to create real or imaginary pets. Now, what you're supposed to do with them and whether

Snapchat is hoping this is just the next Tomogachi remains unclear, but shows just how much experimentation there is going to be in this space, especially among consumer tech. For now, though, that is going to do it for today's AI breakdown. Next up, the main AI breakdown. Before we get to the main part of the episode, one quick note about the AI education beta that you have been hearing about all week. This is the learning program we've been running for a couple months now that's all about tutorials, challenges, basically getting you learning AI by actually using AI faster, more effectively, more efficiently, and hopefully in a way that's more fun too.

If you sign up, you'll get access to a library of 60 plus tutorials, case studies, and challenges, new lessons that drop every weekday, and a passionate community of like-minded learners. We cover everything from LLMs and prompt engineering to image generators, voice synthesizers, AI no cold tools, as well as professional applications like presentation generation website building and more. Due to a bunch of people saying that the window for enrollment was too short, I'm actually going to extend it to Sunday night at 1159 East Coast time. So you have a couple more days to register.

Go to bit.ly slash AI beta to learn more and sign up. I'd love to see you there. But now let's get back to the show. Welcome back to the AI breakdown. It's very clear if you spend any time watching politics or digging into geopolitics especially that artificial intelligence has very quickly become a key dimension of the international conversation. Now, it is a factor both in terms of how countries relate to one another, but also how they

view themselves internally, how they view their internal competitiveness, and how they view their relationship with entrepreneurship and business. Today we're talking about a set of stories that all show how different geopolitical actors are making moves around AI to position themselves in what has become an increasingly important global game. We start here in the U.S. with the announcement of the new National Artificial Intelligence Research Resource. Now, this came out of President Joe Biden's executive order on AI. One of the things that was articulated in that executive order, which, by the way, was I think the longest executive order in U.S. history, was that the United States needed to focus on not only its global competitiveness

relative to the rest of the world, but also making sure that there was equitable access within the U.S. to the infrastructure for participating in the AI revolution. And that is where the N-AIRR comes in. Basically, the National Science Foundation and 10 other agencies, including DARPA, the Defense Advanced Research Projects Agency, NASA, the National Institutes of Health, the National Institutes of Standards and Technology, the National Oceanic and Atmospheric Administration, the Department of Agriculture, the Department of Defense, the Department of Energy, the Department of Veterans Affairs, and the U.S. Patent and Trademark Office

have come together with a set of private sector partners as well to launch this new resource. The 15 private sector partners include many who took that voluntary AI pledge, including Amazon, web services, Anthropic, AMD, Google, Hugging Face, IBM, Intel Meta, Microsoft, Nvidia, OpenAI, and Palantir. So what does the N-A-I-R-R do? Well, it gives people access to AI models, access to compute, access to data sets, software, and training, and makes it available for U.S.-based researchers. The announcement post was called democratizing the future of AI R&D.

The director of the National Science Foundation said in a statement, to continue leading an AI research and development, we must create opportunities across the country to advance AI innovation and strengthen educational opportunities, empowering the nation to shape international standards and igniting economic growth. The NSF blog post continues, The NER pilot will initially support AI research to advance safe, secure, and trustworthy AI, as well as the application of AI to challenges in health care and environment and infrastructure sustainability. The pilot's operations will be organized into four areas of focus.

NAR Open will enable open AI research through access to diverse AI resources via the NER pilot portal and coordinated allocations. NER secure will enable AI research requiring privacy and security preserving resources and will assemble exemplary privacy preserving resources. NER software will facilitate and investigate interoperable use of AI software platforms, tools, and services, and NERPilot resources, and NAR classroom will reach new communities through education, training, user support, and outreach. When you go to the NERPilot webpage, which is at N-A-I-R-R-Pilot.org, it puts the focus as spurring innovation, increasing the diversity of talent, improving capacity, and advancing trustworthy AI. Now, researchers who are interested in this can go on this site, again, NairPilot.org, and participate in a set of

various opportunities. For example, researchers, educators, and students can fill out a survey that hopes to give the organization more information about the challenges that those groups face. They can see a set of pilot resources that include things like pre-trained models, AI-ready data sets, and other relevant platforms, or researchers can apply for computing allocations. Now, when it comes to those computing allocations, the applications are open between now and March 1st of this year. They identify the thematic focus of the allocation call as safe, secure, and trustworthy AI, with research they're interested in around areas like testing, evaluating, verifying, and validating AI systems, increasing the interpretability and privacy of learned models, reducing the vulnerability

of models to families of adversarial attacks, etc. There are six specific compute resource locations that people can apply to or say they don't have a preference. And of course, the question will be how much of a dent these resources make in who has access to actually do advanced research in the AI space. It certainly would seem to be better that it exists than it doesn't, but given just the sheer quantity of compute that is necessary, it'll be interesting to see whether it actually changes anything around the dynamics of this being a very, very top-heavy industry. Now, interestingly, Europe has launched something very similar. Europe has, of course, been better known recently for its attempt to regulate AI, to the chagrin of some leaders in that

region, such as French President Emmanuel Macron, who are worried that Europe is spending so much time on regulating AI, that it will do so at the expense of actually cultivating an AI industry locally. The European Commission is trying to address that with something that they're calling AI factories. This is very similar to this N-AIRR in the U.S. It's meant to give AI startups and researchers in Europe access to better data sets, to more powerful computing resources, and to generally help them actually build. The European Commission tweeted yesterday, With three of the world's five most powerful supercomputers in Europe,

it is time to capitalize on our digital lead and foster trustworthy AI. Our new initiatives include setting up AI factories to support European startups and small and medium enterprises, creating an AI office to oversee the implementation and enforcement of the AI Act, enhancing the availability of language data to boost language models. So once again, the idea of these AI factories from the press release is to, quote, offer a one-stop shop for startups and innovators, supporting the AI startup and research ecosystem in algorithmic development, testing evaluation and validation of large-scale AI models, providing supercomputer-friendly programming facilities and other AI-enabling services. If you take nothing else away from this, it's that compute

is very clearly a political resource now in the same way that oil or energy or anything else like that has been for years. Lastly, today, some interesting comments from the White House's top science advisor. It is no secret that AI has been added to the heap of issues that divide the U.S. and China right now. One only needs to look at the basically ever-expanding sanctions that the U.S. has placed on China, especially when it comes to access to advanced AI chips. But Aradi Prabhakar, the director of the White House Office of Science and Technology Policy, told the Financial Times in a recent interview that despite Chinese-U.S. trade tensions over AI, the countries were planning on trying to work together to lessen AI risks.

She said steps have been taken to engage in that process. We have to try to work with Beijing. F.T. writes, Her comments are an explicit signal that the two powers plan to collaborate on safeguarding the rapidly developing technology, even at a time of heightened trade tensions between the countries. Now, you might remember if you've been listening to this show for a couple months, there was a big question back in November about whether China was going to participate in the UK's Bletchley Park AI Safety Summit. Well, not only did China participate, but they actually signed the Bletchley Park Agreement on standards for the technology. And to add to that,

when Biden and Xi Jinping met together at a summit in California last month, AI was an area that they had committed to work together on. Said Prabhakar, we are at a moment where everyone understands that AI is the most powerful technology. Every country is bracing to use it to build a future that reflects their values. But I think the one place we can all really agree is we want to have a technology base that is safe and effective. So I think that is a good place for collaboration. Now, one other thing you might remember is that the FT had also reported earlier this month that big US AI labs like OpenAI had been engaging in secret meetings with Chinese experts.

on exactly these types of risks. So maybe this is a signal that from the bottoms up, as well as from the top down, hold aside issues of economic competitiveness, AI safety is an area where the world's biggest powers might try, at least, to align. Anyways, always interesting when it comes to AI and geopolitics. I'm sure we'll be seeing lots more of this,

but for now, that is going to do it for the AI breakdown. Until next time, peace.