字幕列表 影片播放 列印英文字幕 Artificial intelligence is already changing business in the United States. Researchers estimate that about 60% of jobs in advanced economies could be impacted by AI. Like with all technological revolutions, I expect there to be significant impact on jobs. I think if this technology goes wrong, it can go quite wrong. Some companies are already seeing the effects. It basically does the job of 700 full time agents, and it's happening probably at an unprecedented pace compared to some other industrial revolutions that we've seen historically. There will be a time in the future where robots can do most of what humans currently do. We're not there yet, and shame on us if we end up using it to make people worse off. Congress so far has focused on research and risk assessment across at least 15 laws on the subject, but experts say there's more work to be done. I've come to the conclusion that it's impossible for Congress to keep up with the speed of technology. Right now, their goal is to basically come up with what's going to be a to do list for future years. As the development of AI speeds up, politicians are looking into robot taxes, 32 hour workweeks, deepfake mitigation and robocalls to confront the looming risks. Will the US government be able to control AI tools? Artificial intelligence is making workers more productive. Productivity comes from improvements in technology. The steam engine was a big technology that that raised living standards. Electricity, computers and now artificial intelligence is the great general purpose technology of our era. For example, one study found that workers who use AI assistants increased their productivity by 14% per hour on average. This particular study focused on customer service call centers. Newer workers, the less skilled workers got the biggest gain. They got up about a 35% gain, whereas the most skilled workers got basically zero gain. So it was kind of a leveler. Part of what seemed to be going on was that it was capturing the tacit knowledge about how to create good answers from the best workers, the best answers, and making it accessible to the rest of the workers. Economists and politicians believe this technology could be deployed very quickly, and in fact, it already has. When we launched this AI chatbot, what blew us away was just that it basically was capable of taking care of two thirds of all the incoming errands that we have over chat, and was doing so with same or even many times higher customer satisfaction score than the human agents were. Sebastian Siemiatkowski is the CEO of Klarna, a Sweden based financial services provider. The company offers buy now, pay later options on websites for over 500,000 merchants, including Nike, Instacart, Yeti, Rei and Brooks Brothers. The first thing the customer writes in the chat is agent like, I want to speak to a human, and I think the reason for that is simply that all of us have had such a horrible experiences of these chatbots that were so not intelligent. But the difference with this chatbot that we launched is that people appreciate it. It's fast. It's to the point it actually resolves their errands in general, an average time of resolution. So like when the customer starts chatting with us and actually resolves, that could be about ten minutes with a human agent while it's been about two minutes with the AI. So I think that unfortunately we may end up having a significant amount of permanent technological unemployment, that income and wealth inequality will rise. The concerns about inequality have led economists to study robot taxes. The idea is to charge companies that use technology to displace workers. The funds would help governments make up for the loss of income and payroll taxes that robots aren't charged, with some revenue set aside to support displaced workers, the ideal level for this tax would be 1 to 3.7%, according to researchers at MIT. By comparison, both employers and employees pay 7.65% for payroll taxes, in addition to other taxes at the federal, state, and local levels. One of the first rules of taxation is whatever you tax, you get less of. Robots are part of what boost technological growth. So I'm not sure we want to have less of that. In 2023, a group of New York State Assembly members introduced a bill that taxes corporations that use AI algorithms that replace workers in the state. But it's not law. Yet lawmakers in Washington are far away from deciding on policy details like this. So I wouldn't expect anything on AI right away. At least not a really impressive comprehensive package. If there is an emergency, say, in the next year, in the next couple of years, that might spur Congress to act. A potential tax on robots would have to be extremely high to alleviate inequality, according to a Kellogg School of Management study. But even then, economists worry that it could shrink the United States economic competitiveness. Other members of Congress take a different approach in light of the rapid development of AI. For example, Bernie Sanders has proposed shortening the workweek to four days. Technology has made working people far more productive, and what has been the result of all of that productivity increase for working people? Almost all of the economic gains of that technological transformation have gone straight to the top, while wages for workers have remained stagnant or even worse. Researchers believe that AI tools are not yet well developed enough to fully replace human beings. If you've used ChatGPT, you know that it's fun and it's cool, but it sometimes makes mistakes. It can hallucinate where it just makes up things that aren't true. That hasn't stopped investors from piling money into the tech sector, expecting huge gains in the future. Companies like Microsoft, Google, Apple, and Meta are all developing artificial intelligence copilots. But the dominant brand, OpenAI, isn't traded on public markets. OpenAI is of the highest quality and gives us the best experience, but there's tons of internal work on top of that amazing technology that they, you know, help us with. Klarna says that OpenAI's assistant tool is estimated to produce $40 million in profit improvement in 2024. Large language models like ChatGPT and Claude and Gemini tend to work better. The bigger they are, the more data they have. It is often kind of a winner take all or winner take most game. Over 50,000 workers lost tech jobs in 2024 as of April. In some cases, leaders were making room for new talent who could develop AI tools. Artificial intelligence is really at play here. When you really look at the numbers of people that are being restructured or being laid off, you're looking at a lot of white collar jobs. Goldman Sachs predicted that 300 million jobs could change as AI tools roll out. Jobs that involve physical work are less likely to be affected. For example, IBM told its employees that it's cutting down jobs in its marketing and communications division, while massively upskilling all of its employees on AI. Revenue per employee is going to be significantly higher than what we've seen historically, and I think it will benefit the employees of those companies because they will also pay more per employee. A survey of 1400 US business leaders suggests that most corporate executives will hire more people as a result of generative AI, not less. Experts believe that's because the tools currently can make workers more productive, but can't replace them altogether. The technology is having an impact in some parts of the market, but I don't think it's having a broad based effect yet. Over the coming years, we'll see more effects. The federal government is in a race against time to understand AI before the products become even stronger. Congress failed to meet the moment on social media. Now we have the obligation to do it on AI before the threats and the risks become real. It's not just about the copyright holders of today and those who will generate new copyright with the assistance of AI. It's about whether America will lead the world. In 2024, the House of Representatives launched a task force to address the wide ranging risks raised by tools that can mimic human thought, language, likeness and more. A part of the idea behind the task force is if you look at who's on it, it's a bunch of different Democrats and Republicans from various committees. The idea here is to kind of have one coordinating body. So as committees begin to do their work, you do have someone who's kind of keeping an eye on everything that's happening and maybe helping to coordinate a little bit as those bills move through committee and, who knows, potentially come to the floor. Leaders in Washington are relying on the industry itself for some of the answers. The White House has garnered commitments from 15 companies to drive safe development of AI. You know, smart regulation can be helpful. It can actually even speed up the adoption of technologies and protect people from harms. But at the same time, overregulating can be harmful and slow down adoption. The US government appears to be moving slower on regulation when compared to its peers. For example, in March 2024, the European Union banned high risk uses of artificial intelligence, which includes applications in critical infrastructure, health care, banking, the justice system and education. The folks in Brussels, they come up with a lot of bureaucratic rules that make it harder for companies to innovate. The entrepreneurial environment isn't there the way it is in the United States. Which means that Americans may be more exposed to the ups and downs of AI as the technology advances. I think what you're going to probably see is what you see for a lot of other issues. You're going to be seeing Republicans taking the route of less regulation, more room for creativity, for growth, and for the economy. And then for Democrats, you're going to see kind of more of a concern. Is AI trying to discriminate against anyone? Could this be something that actually winds up hurting people rather than helping them? The critical question, the one that I would, you know, urge policymakers to consider is how do we support the people whose jobs are affected in a smart way, and so that it becomes truly a positive outcome for the whole of society? There will be a time in the future. Where robots can do most of what humans currently do. We're not there yet. Over time, we're going to have to update our tax system and even move to systems like Universal Basic Income that make sure everybody get a share of the benefits that are created.