The AI Daily Brief: Artificial Intelligence News and Analysis - The FCC Wants AI Robocalls to Be Illegal

Episode Date: February 1, 2024

On this extended brief, NLW covers: the FCC's new push to make AI robocalls illegal; VW's new AI lab, Mastercard's fraud detection AI, and much more. 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, the FCC is making AI-generated robocalls illegal. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown on Network for more information about our YouTube, our Discord, and our newsletter. Hello, friends, quick note before we dive in. I mentioned that this week was a little weird because of some things going on that are exciting and have to do with future plans that you will learn about in a couple of weeks. But today's episode is impacted by that, and so instead of our normal format of a brief, followed by a main episode. Today, we just have an extended brief. It won't be too short,

and I extended it so that we covered even more of the headline news than we normally do, but that's why this episode is a little bit different. Back to normal tomorrow. Welcome back to the AI breakdown brief. All the AI headline news you need in around five minutes, although today's is an extended brief, so it will be a little bit over that five-minute mark. There are a bunch of big themes that already, just one month into 2024, we keep coming back to. One that I talk about a lot is this idea of integration and how AI is not just about pushing the state of the art, but also about coming into the tools that we use day in and day out. But then there are also some big society level themes.

And of course, with the U.S. presidential elections coming up this year, AI's impact on that process is going to be very, very front and center. One of the big stories on that front over the past couple weeks was a set of robocalls that made it to New Hampshire Democrats from an AI-generated Biden encouraging them not to come out vote in a primary and instead save their vote for November when, according to AI Biden, it really mattered. Well, now the FCC is trying to make it so that those calls are not just annoying, but actually illegal. So this is actually a weirdly nuanced thing. As TechCrunch points out, voters' suppression in a place like New Hampshire is already illegal. So if New Hampshire authorities

could figure out who had actually made these calls, they could be charged with voter suppression. However, if the actual use of voice cloning technology in automated calls is illegal itself as well, not just the fact that it's deployed for voter suppression purposes, it gives authorities more tools to go after that type of behavior. Said FCC chairwoman Jessica Rosen-Worsel, that's why the FCC is taking steps to recognize this emerging technology as illegal under existing law, giving our partners at state attorneys general's offices around the country new tools they can use to crack down on these scams and protect consumers. Now, this really is such a weirdly specific thing that shows just how nuanced and gray area and complex, figuring out how AI interacts with existing laws really will be.

For example, here's a quote from TechCrunch. The FCC already uses the Telephone Consumer Protection Act as the basis for charging robocallers and other telephone scammers. The TCPA already prohibits artificial voices, but it is not clear that cloned voices fall under that category. It's arguable, for instance, that a company could use the generated voice of its CEO for legitimate business purposes. Now, interestingly, under new rules, even that legitimate use case would now become illegal, and so the question becomes, do we want to focus rules around preventing the negative or enabling the positive? As TechCrunch puts it, the fact is that legal applications of the tech are fewer in number and less immediately important than the illegal applications. Therefore, the FCC proposes

to issue a declaratory ruling that AI-powered voice cloning causes a call to fall under the artificial heading. Now, in terms of process, this proposal will now be spread internally and voted upon, and will only become public when it is officially adopted. Next up, staying on a U.S. policy issue but moving to a very different dimension, the United States has, of course, been increasing restrictions on what U.S.-based companies can and can't sell to China when it comes to AI hardware. And once again, NVIDIA has tried to create a chip that falls underneath the latest round of export restrictions.

According to Reuters, Invidia has started taking pre-orders for a new China-specific chip, which they're calling the H-20. The H-20 is the most powerful of three different, in Nvidia chips that the company has been developing for the Chinese market after the recent extended bands, and it appears that they're trying to price it similarly to a rival product from Huawei. Now, sources say that while it's priced similarly to Huawei's ascend 910b chip, the H20 might actually be less powerful. Reuters writes, in terms of specifications, one example of where the H20 appears to lag the 910B in its PF-32 performance, a critical

metric that measures how quickly a chip can process common tasks, and which is rated at less than half of its rival's capacity. However, the H20 appears to have an advantage over the 910b in terms of interconnect speed, which measures how quickly data can transfer between chips. That means the H20 remains competitive with the 910B in applications that require linking a large number of chips together to work as a system. Now, of course, part of why Chinese companies have been more inclined to increasingly choose Huawei products is that no one is particularly confident that this will be the last round of restrictions, and so it may be that just as they adopt this new H20 standard, all of a sudden in a year, that's no longer available either. Now, speaking of the U.S.-China AI battle, the Pentagon has

added another Chinese chipmaker to a roster of companies that it accuses of aiding China's military. The companies in question are the Yangtze Memory Technologies Coe and facial recognition for Megvi, who have now been added to this list, which is called the Section 260H. writes Bloomberg. The U.S. government has in past years grown increasingly concerned about what it calls China's military civil fusion strategy, describing the concept as regular civilian companies aiding the people's liberation army in some fashion. Chinese foreign ministry spokesman, Wang Wenbin told reporters, the U.S. move violates market competition principles and international trade rules, undermines foreign businesses' confidence, undermines the interest of companies and investors, and will backfire.

Now, moving over to a different geopolitical area, a new report commissioned by Amazon Web Services finds that a third of European businesses adopted some form of AI in 2023, said Tunisia Randari, the managing director at AWS, last year was pivotal. From Mamo to Milan, the accessibility of generative AI created an appetite for experimentation among consumers and businesses. The accelerated uptake of AI last year has helped put Europe on track to meet its digital decade goals. The report also estimated that the positive economic impact of AI would be something on the order of 3.4 trillion euros. Now, one of those companies that is adopting AI is Volkswagen.

The Carmaker has announced that it has founded a new artificial intelligence lab to generate new product ideas, including potential partnerships with tech companies in China, North America, and Europe. Says Reuters, the aim is to generate early stage prototypes in areas like AI optimized charging cycles, predicted maintenance services, and voice recognition. For me, right now, the notable thing is that you can still get pressed just by saying you're starting an AI lab without giving any more details. Another company touting its new AI integrations is MasterC. The company told CNBC that it's launching a new generative AI model that is focused on better detection of fraud.

MasterCard claims that it can help institutions improve fraud detection rates by as much as 300%. Said AJ Bala, MasterCard's president of cyber and intelligence business unit, we're using the transformer models which basically help get the power of generative AI. It's all built in house and we've got all kinds of data from the ecosystem. Because of the very nature of the business we are in, we see all the transaction data which comes to us from the ecosystem. The proprietary algorithm is trained on data from the roughly 125 billion transactions that go through MasterCard's network annually. MasterCard's big competitor Visa is of course also investing in AI, announcing, for example, back in October a new $100 million

venture fund for generative AI startups. Now, according to one report, MasterCard and Visa are not alone in taking advantage of generative AI in the banking sector. Indeed, a new report published today from the Burning Glass Institute, which is a nonprofit research center, and SHRM, which was formerly the Society for Human Resource Management, says that banking is one of the sectors that's most likely to be affected by generative AI. From the New York Times, the research estimates that banks and some tech companies spend 60 to 80 percent of their payrolls or more on workers and occupations most likely to be affected by the new technology. Now, given this scope of disruption, it's no surprise that AI has been, as Yahoo Finance puts it, front and center on quarterly earnings calls.

Reuters analysis said that AI or artificial intelligence has been said on 38% of calls held by S&P 500 companies in January, which is up from 34% on the last round of earnings calls last quarter. On the average S&P 500 call this January, AI has been mentioned 3.5 times. Now, one cool tool announcement comes from Shopify. The company has introduced a new magic AI image editor that can help make product pictures from people on its e-commerce platform look more professional. Writes the Verge. Glenn Coates, Shopify's vice president of product, says many sellers on the platform don't have

access to big studios or professional photographers, but they still need to set their items up for success. Said Glenn, we know not everyone has access to studios or mannequins, so we want to lower the barrier to entry for entrepreneurs to present their products in the best light. We're not forcing anyone to use this feature, but it will be an extra helping hand. Now, this is firmly in that integration theme that I keep coming back to, and I think that this is going to be a super useful application of AI for a huge number of people who are already using Shopify's tools. Now, for those who are worried that maybe their customers won't be interested in AI-generated

media, a new survey of 700 U.S. consumers found that AI-generated content was actually preferred by humans more than human-generated content. Basically, these humans got six battles. The battles included writing social media ads, writing a blog post paragraph, writing a social media post, and writing a product description. In each of the six cases, A.I. was more preferred to human copy. Why that was? Well, there were three standouts. People liked that the AI generated content quickly got to the point, that it clearly highlighted value propositions, and that it was easier to read and understand. This will be bad news for copywriters and good news for the platforms that are increasingly investing in AI-powered tools for ad generation.

Now, speaking of big pattern changes, The Guardian reports that more than half of UK undergrad say they use AI to help with essays. A survey of more than 1,000. thousand UK undergrads conducted by the Higher Education Policy Institute found 53% were using AI to generate material for work they would be scored on. My interpretation of this is that that means that 47% of surveyed UK undergraduates were lying about their use of AI to help them on essays. Now remember, this doesn't mean that they're copy and pasting directly. Indeed, just 5% admitted to copying and pasting unedited AI generated text into their assessments. These are students who are adopting AI as part of their workflow, and that, my friends, is completely an

I mean completely inevitable. Hopefully as we start to recognize that this year, we'll focus more on how people can use AI well rather than trying to convince them not to use it. Overall, AI continues to be the big topic. Fox News recently hosted former U.S. presidential candidate Andrew Yang, who warned that the U.S. was not doing enough to prepare for AI's impact and its dramatic changes. Of course, I think that if all of them were listening to the AI breakdown and chatting with all of you in our Discord, they'd probably be a heck of a lot more prepared. However, for now, that is going to do it for today's extended brief. I appreciate you listening or watching as always. And until next time, peace.