 |
|
|
|||||
If you listen to the CEOs of elite AI companies or take even a passing glance at the U.S. economy, its abundantly obvious that AI excitement is everywhere. Americas biggest tech companies have spent over $100 billion on AI so far this year, and Deutsche Bank reports that AI spending is the only thing keeping the United States out of a recession. Yet if you look at the average non-tech company, AI is nowhere to be found. Goldman Sachs reports that only 14% of large companies have deployed AI in a meaningful way. What gives? If AI is really such a big deal, why is there a multi-billion-dollar mismatch between excitement over AI and the techs actual boots-on-the-ground impact? A new study from Stanford University provides a clear answer. The study reveals that theres a right and wrong way to use AI at work. And a distressing number of companies are doing it all wrong. What can AI do for you? The study, conducted by Stanfords Institute for Human-Centered AI and Digital Economy Lab and currently available as a pre-print, looks at the daily habits of 1,500 American workers across 104 different professions. Specifically, it analyzes the individual things that workers actually spend their time doing. The study is surprisingly comprehensive, looking at jobs ranging from computer engineers to cafeteria cooks. The researchers essentially asked workers what tasks theyd like AI to take off their plates, and which ones theyd rather do themselves. Simultaneously, the researchers analyzed which tasks AI can actually do, and which remain out of the technologys reach. With these two datasets, the researchers then created a ranking system. They labeled tasks as Green Light Zone if workers wanted them automated and AI was up to the job, Red Light Zone if AI could do the work but people would rather do it themselves, and Yellow Light (technically R&D Opportunity Zone, but Im calling it Yellow Light because the metaphor deserves extending) if people wanted the task automated but AI isnt there yet. They also created whats essentially a No Light zone for tasks that AI is bad at, and that people dont want it to do anyway. The boring bits The results are striking. Workers overwhelmingly want AI to automate away the boring bits of their jobs. Stanfords study finds that 69.4% of workers want AI to free up time for higher value work and 46.6% would like it to take over repetitive tasks. Checking records for errors, making appointments with clients, and doing data entry were some of the tasks workers considered most ripe for AIs help. Importantly, most workers say they wanted to collaborate with AI, not have it fully automate their work. While 45.2% want an equal partnership between workers and AI, a further 35.6% want AI to work primarily on its own, but still seek human oversight at critical junctures. Basically, workers want AI to take away the boring bits of their jobs, while leaving the interesting or compelling tasks to them. A chef, for example, would probably love for AI to help with coordinating deliveries from their suppliers or messaging diners to remind them of an upcoming reservation. When it comes to actually cooking food, though, theyd want to be the one pounding the piccata or piping the pastry cream. The wrong way So far, nothing about the studys conclusions feel especially surprising. Of course workers would like a computer to do their drudge work for them! The studys most interesting conclusion, though, isnt about workers preferencesits about how companies are actually meeting (or more accurately, failing to meet) those preferences today. Armed with their zones and information on how workers want to use AI, the researchers set about analyzing the AI-powered tools that emerging companies are bringing to market today, using a dataset from Y Combinator, a storied Silicon Valley tech accelerator. In essence, they found that AI companies are using AI all wrong. Fully 41% of AI tools, the researchers found, focus on either Red Light or No Light zone tasksthe ones that workers want to do themselves, or simply dont care much about in the first place. Lots more tools try to solve problems in the Yellow Light Zonethings like preparing departmental budgets or prototyping new product designsthat workers would like to hand off to AI, but that AI still sucks at doing. Only a small minority of todays AI products fall into the coveted Green Light zonetasks that AI is good at doing and that workers actually want done. And while many of todays leading AI companies are focused on removing humans from the equation, most humans would rather stay at least somewhat involved in their daily toil. AI companies, in other words, are focusing on the wrong things. Theyre either solving problems no one wants solved, or using AI for tasks that it cant yet do. Its no wonder, then, that AI adoption at big companies is so low. The tools available to them are whizzy and neat. But they dont solve the actual problems their workers face. How to use AI well For both workers and business leaders, Stanfords study holds several important lessons about the right way to use AI at work. Firstly, AI works best when you use it to automate the dull, repetitive, mind-numbing parts of your job. Sometimes doing this requires a totally new tool. But in many cases, it just requires an attitude shift. A recent episode of NPRs Planet Money podcast references a study where two groups of paralegals were given access to the same AI tool. The first group was asked to use the tool tobecome more productive, while the second group was asked to use it to do the parts of your job that you hate. The first group barely adopted the AI tool at all. The second group of paralegals, though, flourished. They became dramatically more productive, even taking on work that would previously have required a law degree. In other words, when it comes to adopting AI, instructions and intentions matter. If you try to use AI to replace your entire job, youll probably fail. But if you instead focus specifically on using AI to automate away the parts of your job that you hate (basically, the Green Light tasks in the Stanford researchers rubric), youll thrive and find yourself using AI for way more things. In the same vein, the Stanford study reveals that most workers would rather collaborate with an AI than hand off work entirely. Thats telling. Lots of todays AI startups are focusing on agents that perform work autonomously. The Stanford research suggests that this may be the wrong approach. Rather than trying to achieve full autonomy, the researchers suggest we should focus on partnering with AI and using it to enhance our work, perhaps accepting that a human will always need to be in the loop. In many ways, thats freeing. AI is already good enough to perform many complex tasks with human oversight. If we accept that humans will need to stay involved, we can start using AI for complex things today, rather than waiting for artificial general intelligence (AGI) or some imagined, perfect future technology to arrive. Finally, the study suggests that there are huge opportunities for AI companies to solve real-world problems and make a fortune doing it, provided that they focus on the right problems. Diagnosing medical conditions with AI, for example, is cool. Building a tool to do this will probably get you heaps of VC money. But doctors may not wantand more pointedly, may never usean AI that performs diagnostic work. Instead, Stanfords study suggests theyd be more likely to use AI that does mundane thingstranscribing their patient notes, summarizing medical records, checking their prescriptions for medicine interactions, scheduling followup visits, and the like. Automate the boring stuff is hardly a compelling rallying cry for todays elite AI startups. But its the approach thats most likely to make them boatloads of money in the long term. Overall, then, the Stanford study is extremely encouraging. On the one hand, the mismatch between AI investment and AI adoption is disheartening. Is it all just hype? Are we in the middle of the mother of all bubbles? Stanfords study suggests the answer is no. The lack of AI adoption is an opportunity, not a structural flaw of the tech. AI indeed has massive potential to genuinely improve the quality of work, turbocharge productivity, and make workers happier. Its not that the tech is overhypedweve just been using it wrong.
Category:
E-Commerce
The rise of artificial intelligence in recent years, along with the surge in AI-generated online content, has given more credibility to a decades-old conspiracy theory known as the Dead Internet Theory. It holds that most of the content we encounter online isnt actually produced by living humans but by lifeless bots. AI is increasingly turning the once-fringe theory into a reality, but even today, at least one of the participantsthe living, breathing observer browsing the web on the other side of the screenis still usually a real, sentient being. Yet this may not be true for much longer. Thanks to AI systems increasing reliance on a technology known as headless browsing, artificial intelligence is becoming a primary consumer of the internet. And if that happens at scale, the internet will truly be a land of the unliving. Heres what you need to know about headless browsing, a term youll likely hear increasingly often in the years ahead. Headless browsing is nothing new Nearly every web browser youve ever usedwhether its Google Chrome, Apples Safari, or even Microsofts old Internet Exploreris a traditional visual browser. It features a graphical user interface (GUI), which includes buttons, tabs, scrollbars, and, of course, a large window that displays content (i.e., a website) that you can see. You navigate a visual browser mainly by clicking with your mouse cursor on hyperlinks or other buttons on a web page. If the web page requires you to enter text, such as in a form, you click in the text entry field and use the keyboard to type your characters. But for decades, another kind of browser has existed: the headless browser. A headless browser has no graphical user interface (GUI). It has no window that displays a webpage and does not support pointing and clicking with a mouse cursor. Instead, a headless browser processes a website’s content by reading its code directly. It interacts with the site, such as clicking on a link to go to the next page or entering text into a form, all through direct interaction with its code. Since humans are visual creatures, its clear why GUI browsers are the primary way most of the world accesses the internet. So then, what are headless browsers used for? Historically, they have been tools for web developers, as enterprise proxy provider Oxylabs explains. Because every graphical user interface element on a webpage has corresponding code, an automated program designed to help devs find errors on a website running through a headless browser can interact with that website just like a person wouldbut much faster since no visual interface needs to be displayed. The traditional benefit of headless browsers is that websites become more stable and reliable because headless browsing allows errors to be found relatively quickly. But human developers arent the only ones using headless browsers anymore. Headless browsing in the age of AI Once a tool for web developers and other programmers, headless browsers are now being employed by new userswho dont have heads at all. Increasingly, artificial intelligence systems are the primary users of headless browsers. AI browsers, such as Perplexitys Comet, use headless browsing to scan websites to carry out your prompts quickly. For example, when you prompt an AI browser for a list of the capitals of the 50 United States, the browsers AI will read the content of numerous websites via headless browsing to quickly retrieve the answer. But headless browsing goes beyond letting AI scan a website to retrieve information. As artificial intelligence systems evolve from being simple answer bots to becoming personal assistantsknown as AI agentsheadless browsing is also being utilized by these agents to interact with websites on your behalf, performing tasks like clicking links, checking boxes, or even adding items to your shopping cart. A large part of why an AI agent can perform tasks you prompt it to do so quickly is due to headless browsing. For example, say you prompt an AI browser to order the ingredients you need to make Thanksgiving dinner from multiple grocery websites. The browsers AI agent isnt actually perusing grocers websites through any visual interface and then clicking on Buy Now buttons to find and add items to your shopping cart. Its using headless browsing to read and interact with the websites code directly. But while headless browsing makes AI more efficient and versatile, theres a negative side to AIs use of it, particularly if youre a website or one of its advertisers. An internet where AI is the main user, not humans As more people turn to agentic AI and AI browsers, these AI systems will utilize headless browsing to visit websites and carry out tasks assigned by humans. This means that AI has the potential to be the primary type of user that is visiting a website. And there are already signs of this happening. A report from the AI monetization platform TollBit last month showed that, for the most recent quarter, human traffic to the websites that TollBit monitored declined by 9.4%, while AI traffic continued to rise. And its rising a lot. In the first quarter, TollBit found that 1 out of every 200 visitors to the sites it monitors was AI. By the second quarter, AI visitors accounted for 1 out of every 50 visitors. Thats a fourfold increase in less than a year. TollBits report goes on to note that when AI agents visit a website, the website often has no way to tell that it’s an AI and not a human being. Thats terrible news for companies, which rely on web advertising to pay the bills. Advertisers sell things to human beings, and if advertisers can no longer trust whether a website knows precisely how many actual people are visiting it, they likely arent going to spend their limited ad dollars on that site. For what its worth, an executive at an unnamed large digital publisher told Digiday that they believed headless browsing does not currently pose a major issue for publishers. However, they noted that if big players in the AI space, such as OpenAI or Google, adopt the technology for their AI agents, headless browsing could become a significant concern. And if headless browsing does become the norm, it also means that the Dead Internet Theory could take on an expanded meaning. No longer would the phrase be used to signify only an internet where human beings do not make the majority of the contentbut where the majority of the browsing is no longer done by humans either.
Category:
E-Commerce
If you ride along a bike path in the U.K. city of Leeds and approach a street, the traffic light can automatically turn green for youor stay green if you’re already midway across. The city is one of a growing number testing technology that uses sensors, anonymous data, and AI to make it easier to cross streets. Made by a company called VivaCity (known as Viva in the U.S.), the sensors can detect cyclists and pedestrians from more than 200 feet away. In some cases, someone on a bike might not need to stop at the corner. Pedestrians can keep walking without breaking their stride. (Cities can choose to program traffic signals to give cyclists either a shorter wait or full priority.) Traditionally, most traffic signals force people who arent in cars to push a beg button and waitor risk their life to cross while the lights still red. If automated sensors exist, theyre typically just for cars. [Image: Viva] At the moment, a lot of traffic signals don’t detect cyclists, says Matt Shaw, head of product at VivaCity. If they’re really basic, they operate on a fixed time schedule, so it will just rotate 30 seconds at a time. Or they might have vehicle detection, so they know if a car’s approaching, but they don’t know if a cyclist is. Wires embedded in the pavement can detect metal, but often miss bikes. VivaCitys technology also analyzes direction, so the automatic walk sign isnt triggered if a pedestrian is just passing by without intending to cross. It also counts the number of people waiting, so cities can choose to use a formula to change the light faster if more people are waiting. Most traffic controllers now have no idea if it’s one pedestrian or a hundred, says Shaw. If you’re in New York City and somebody pushes the button, you’ve got no sense of how many people are waiting. [Image: Viva] Unlike standard traffic signals, Viva’s system also knows if someone in a wheelchair or an elderly person with a walker is still crossing. Being able to know if a pedestrians still on the road, and hold the green light for them, is pretty important, he says. (The data collection focuses on privacy; after the AI analyzes the video feed, it deletes it, leaving only the number of people and the path that theyre taking, not their identity. The data cant be used for enforcement.) In Leeds, the city hasnt yet gone as far as giving cyclists and pedestrians full priority at major intersections. But at certain crossings where bike paths or sidewalks meet a road, the sensors already prioritize people walking or biking. The tech is one piece of a bigger strategy to become a city where you dont need a car, as the city puts it. That also includes improving bus service and building a better network of bike routes and bike parking. Making streets easier to crossand shrinking the time that cyclists or pedestrians waithelps make it a little more likely that people will want to walk or bike. Some cities are using the underlying data without yet connecting to traffic controllers. In New York City, for example, the Department of Transportation has been using the sensors at some intersections to track trends over time, from the number of bikes or scooters to how fast theyre traveling and the paths that people take to cross the street. The technology can also track near misses, which lets cities flag dangerous intersections and design interventions, such as changing the timing of signals or banning turns on red at intersections where cyclists have repeated close calls with turning vehicles. You cant solve the problem if you dont understand where people are cycling, Shaw says.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||