The AI Daily Brief: Artificial Intelligence News and Analysis - Work Disruption? Klarna's AI Assistant Does Work of 700 Full Time Agents

Episode Date: February 28, 2024

Klarna's new AI chatbot, powered by OpenAI, has showcased impressive capabilities in its first month, effectively performing the tasks of 700 full-time customer service agents. Engaging in 2.3 million conversations, the bot handled two-thirds of Klarna's customer service interactions, achieving satisfaction rates on par with human agents. Plus Apple ditches car plans to focus on AI

Today on the AI breakdown, Apple has killed their electric car plans and refocused on AI. Before that on the brief, a fintech company says that its open AI powered chatbot is doing the work of 700 customer service agents. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.com. Network for more information about our Discord, our newsletter, and our YouTube. Welcome back to the AI breakdown brief, all the AI headline news you need in around five minutes. One of the big themes that people are constantly looking out for when it comes to discussions of AI is how it's going to impact jobs, or more specifically, evidence that it is impacting jobs in some specific way. Today, then, it will not be surprising that there has been a ton of buzz in conversation about fintech company Klarna

discussing the results of their first month of using an AI assistant that is powered by OpenAI. The headliner stats is that in this month, the Klarna bot had 2.3 million conversations, which represented around two-thirds of Klarna's customer service chats. In terms of customer satisfaction, its score is on par with human agents. Klarna also reports that it's more accurate in its resolution, with a 25% drop in repeat inquiries. They also report that it's faster, with customers resolving their issues in less than two minutes compared to 11 minutes previously. But the big banner headline was that they said it's doing the equivalent work of 700 full-time agents.

Now, Klarna got into some hot water back in 2022 when they laid off around 10% of their staff. At the time, the company cited issues like economic uncertainty, inflation, the likelihood of a recession, and obviously since then, we've seen lots of other tech companies go through a similar process of staff reduction. Now, one of the things that some have noticed is that the number of people who were laid off back then is roughly similar to the 700 jobs figure that they now say this AI assistant is doing the work of. However, in a statement to Fast Company, they said, this is in no way connected to the workforce reductions in May 2022, and making that conclusion would be incorrect. We chose to share the figure of 700 to indicate the more long

term consequences of AI technology, where we believe it is important to be transparent in order to create an understanding in society. We think it's important to proactively address these issues and encourage a thoughtful discussion around how society can meet and navigate this transformation. There are a number of different takes on this. First of all, many are pointing out that this suggests that AI is not just some future thing, but is here now and is going to have a big impact. Gavin Baker writes, AI is going to have a massive impact on every industry. Others didn't really buy it. Market commentator Sean Tuffy shared a meme basically intimating that Klarna was making up the number when it said that it thought that the assistant thought was going to drive $40 million in additional profits.

Now, it is worth noting that Klarna is on the IPO path and that could be coming soon, and so showing something which investors are going to read as capital efficiency could be part and parcel in increasing interest around that IPO. For his part, the CEO of Klarna, Sebastian Semiadkowski, repeated the message that we heard in that previous statement. He tweeted today, continuing the discussion around AI like this is important to me. While good to acknowledge the positive benefits, it's important to bring transparency around the wider societal impact so we can navigate the transition into the era of generative AI responsibly. Now, another reaction came from markets who looked at companies that provide call center jobs with a new skeptical eye. For example,

French company teleperformance, which is a call center business, saw their shares drop as much as 29% after the Klarna announcement. A Morgan Stanley analyst wrote, the AI debate has raged on for teleperformance in 2024. Until questions around impact of pricing deflation, automated volume, and shape of future earnings are answered, the stock performance may remain challenged. Now, teleperformance was perhaps most impacted because it actually supplies these services to Klarna, but other publicly traded call center companies like U.S. Concentrics corp and Task Us Inc also were down significantly in pre-market trading. Concentric, for example, was down 12%. It is useful to note that there was also some skepticism of this when people actually went and tried the

bot themselves. FinTech content creator Simon Taylor writes, I tried the chatbot. It's kind of shit. Did Klarna get great PR because they did a chat bot here? Gerger Lio Rose writes, when something sounds too good to be true, maybe it is. Did Klarna really fully replace 700 customer support agents overnight? I did what few retweeting this story did, tried Klarna's AI assistant. It's underwhelming. It recites exact docs and passes me on to human support fast. I think whatever the truth is when it comes to Klarna specifically, the fact that this story has caught on like wildfire tells you a lot about the state of the AI discourse right now, and frankly the state of AI fears.

Now, continuing to follow up on a couple stories from earlier this week, the pressure on Google around the Gemini controversy seems to be continuing to mount, and we've finally gotten some statements from Google CEO Sundar Pichai in the form of an internal company memo. Pirate Wires published the entire piece, and it reads, Hi, everyone, I want to address the recent issues with problematic text and image responses in the Gemini app formerly barred. I know that some of its responses have offended our users and shown bias.

To be clear, that's completely unacceptable and we got it wrong. Our teams have been working around the clock to address these issues. We're already seeing a substantial improvement on a wide range of prompts. No AI is perfect, especially at this emerging stage of the industry's development, but we know the bar is high for us and we will keep at it for however long it takes. And we'll review what happened and make sure we fix it at scale. Our mission to organize the world's information and make it universally accessible and useful is sacrosanct. We've always sought to give users helpful, accurate, and unbiased information in our products.

That's why people trust them. This has to be our approach for all our products, including our emerging AI products. We'll be driving a clear set of actions including structural changes, updated product guidelines, improved launch processes, robust evals and red teaming, and technical recommendations. We are looking across all of this and we will make the necessary changes. Even as we learned from what went wrong here, we should also build on the product and technical announcements we've made in AI over the last several weeks. That includes some foundational advances in our underlying model, e.g. our 1 million long

context window breakthrough and our open models, both of which have been well received. We know what it takes to create great products that are used and beloved by millions of people and businesses, and with our infrastructure and research expertise, we have an incredible springboard for the AI wave. Let's focus on what matters most, building helpful products that are deserving of our users' trust. As some have noted, what was not included there was any sort of timeline for getting Google human image generation back online. Another reported story from Google is that according to AdWeek, the company has been quietly striking deals with some publishers to use their new generative AI tools to actually write and publish stories. Writes End Gadget, the deals reportedly

worth tens of thousands of dollars a year are apparently part of the Google News Initiative, a six-year-old program that funds media literacy projects, fact-checking tools, and other resources for newsrooms. AdWeek writes, the beta tools let under-resourced publishers create aggregated content more efficiently by indexing recently published reports generated by other organizations like government agencies and neighboring news outlets, and then summarizing them and publishing them as a new article. Now, while right now everyone is very hopped up to jump down Google's throat on anything, the idea that AI could be used, particularly to support smaller or more local news organizations is something that a lot of people are positively enthusiastic about.

However, given that these are just reports and that we don't really know what's going on, people are tending to focus in on some of the bad things. For example, and gadget rights, of note, publishers in the program are apparently not required to disclose their use of AI, nor are the aggregated websites informed that their content is being used to create AI written stories on other sites. Lastly, today, one more update on the Mistral Microsoft story. It was a smaller part of the announcement, but people were wondering how much Microsoft had actually invested in Mistral as part of this new partnership, and it turns out it's a very small amount, around $16 million. Mistral said that this is an extension of the Series A and doesn't change the valuation of the company. Basically, this leaves Microsoft owning less than 1% of the company.

Now, of course, the bigger questions around the Mistral Microsoft deal tend to be larger implications for the relationships of big tech to startups, what it means for open source now that Mistral Large is available proprietary only through Azure. there are also questions around EU AI Act lobbying, but that will be the subject for another show. For now, that's going to do it for the AI breakdown brief. Up next, the main AI breakdown. Welcome back to the AI breakdown. Big news out of Cupertino, California today, as Apple has apparently canceled its electric car project, a decade-long effort and is shifting resources to mainstreaming its Vision Pro headset, as well as, you guessed it, artificial intelligence. So for those who haven't been paying attention to Apple's electric car efforts, the initiative was known as Project Titan. And it came from the idea that effectively the car is the ultimate mobile product.

If Apple was in the business of mobile products, then maybe an electric vehicle could be its next big thing. Part of it was driven also by a concern that at some point the phone market was going to be saturated and was going to be able to account for less and less growth. Globally speaking, the iPhone still makes up about half of Apple's sales. Now, Project Titan was never actually confirmed by Apple. And the news that they've canceled this effort, which was reported first by Bloomberg, has not been confirmed either. There are a bunch of dimensions of this story which are really interesting. Many of them are inside Apple and what it means for their generative AI efforts, which we'll get into in just a moment. But there is also part of the story which we have to

tell, which has to do with the struggling EV industry. The Financial Times, for example, writes, the cancellation comes as the EV industry has slowed in recent months. In January, Tesla warned that flagging demand, high interest rates, and intensifying competition would lead to slower sales growth rates this year. Ford and General Motors have paused plans to expand their EV manufacturing in recent months, and last week, Rivian announced that it would cut 10% of its workforce and keep deliveries flat this year. So where did this reporting come from if it wasn't confirmed by Apple? It was, of course, Bloomberg's Mark Germann, who is one of the industry's best-known Apple Watchers. German wrote yesterday, Apple is canceling a decade-long effort to build an electric car

according to people with knowledge of the matter, abandoning one of the most ambitious projects in the history of the company. Apple made the disclosure internally Tuesday, surprising the nearly 2,000 employees working on the project. The decision was shared by Chief Operating Officer Jeff Williams and Kevin Lynch, a vice president in charge of the effort. The two executives told staffers that the project will begin winding down and that many employees on the car team, known as the Special Projects Group, or SPG, will be shifted to the Artificial Intelligence Division. Those employees will focus on generative AI projects and increasingly key priority for the company. Elon Musk, of course, the CEO of Tesla, retweeted the news with a salute emoji and a cigarette emoji, perhaps conveying

a recognition of how hard this market is. Now, when it comes to the question of investor reactions, one might have thought that this would be seen as a sign of weakness. Apple is not generally in the business of canceling big ambitious projects. However, the reaction has been exactly the opposite. it. Shares were up about 1% after the news broke yesterday, and many analysts were unequivocal in their support for the decision. Wed Bush analyst Ann Ives writes, Apple has canceled its decade-long effort to build an EV. The writing was on the wall for Apple with a much different EV landscape forming that would have made this an uphill battle. Most of these Project Titan engineers now all focused on AI at Apple, which is the right move. He also said that it's a strategic move that

they applaud loudly. Baker Avenue wealth management's King Lip said, as an Apple shareholder, it's music to our ears. What do you make of this report? Market seems to like it. Do you? Hi, Scott. Yeah, I would say as an Apple shareholder, it's music to our ears, frankly, that the company is terminating the car project. We were honestly never big fans of the car project.

Building a car is capital intensive. It's a low margin. They're serious entrenched players already in the EV market. So now that the company's, you know, officially dedicating more resources. It's something that as a shareholder, we're always kind of wondering how committed is Apple to their generative AI. And it looks like they're committing some serious resources to it. Let me ask you this. Given the installed base, I mean, we're pretty much a product company first and foremost, software. I get it. How big do they need to be an AI to

to make a shareholder like you happy? Well, you know, first of all, I think a lot of degenerative AI is going to be built into their operating system, into the Macs. It's going to be built into the iPhone software, for example. And as we know, there's a couple billion devices, active devices that Apple has. So the consumer distribution for Apple is probably second to none, you know, from that perspective. So I think there's a lot of opportunities there. So in that little clip, you really have the two sides of this discussion. On the one hand, a concern from shareholders about the EV market and how good it was to enter in the first place. and on the other, a concern that has been lingering that Apple wasn't focused enough on generative

AI. Now, of course, this has been something that we've talked about numerous times here, that Apple's absence from this space, as really the only big tech company to be absent from the space, has been extremely notable. The market has widely been willing to give them a pass, given Apple's propensity to do things not first but best, but even that was having its limits, and I think the response of the market to this news, is indication that people were starting to get a little bit concerned around when Apple's AI strategy was really going to ramp up. Now, in a further analysis piece, Mark German again from Bloomberg wrote, Apple's future isn't going to hinge on selling $100,000 cars with self-driving features.

Instead, it will focus on catching up with rivals in the generative AI industry, where chatbots from OpenAI and Google have captured the imagination of consumers and investors. Bloomberg intelligence analysts said shifting resources towards generative AI is the right call, quote, given the long-term profitability potential of AI revenue streams versus cars. German also points out that self-driving car efforts are AI efforts, something that Elon Musk has talked about frequently. He writes, Apple's decade-long car effort was also an AI challenge in itself. Apple attempted to build an artificial intelligence system that was powerful and energy-efficient enough to make a car fully autonomous. With that off the table, Apple can concentrate on applying AI to its current products, including the iPhone and the iPad, and avoid falling further behind tech peers.

So when should we start to actually hear more about what Apple is going to do in AI? there are a couple moments that people are watching for this year. First is that its WWDC conference will happen in June, and it's anticipated that Apple will reveal a set of new AI capabilities for their software there, and it is also widely reported that the first AI features will be rolled out in iOS 18, which is slated to be released around September with the next generation of the iPhone. This will be, of course, a jarring shift for those 2000 engineers who have been working on this project for years. But as someone in the AI space who's excited to see what Apple can do,

and how much they can bring actually useful integrated tools into the platforms that we already use, I can't say I'm not excited to see how this supercharges their efforts. For now, though, that is going to do it for today's AI breakdown. Until next time, peace.