engineers to spend more money than a finance analyst,” says the boss of one software firm. “But anyone within finance, it shouldn’t be zero.” But usage (also known as input) measures don’t tell you whether any value is being created. And even when AI is producing benefits, points out Mika Ruokonen of LUT University in Finland, lots of them accrue quietly to individuals in the form of scattered time savings. If an extra hour or two of an employee’s time has been freed up to shop on Vinted, someone has benefited (mainly Vinted) but it isn’t your company (unless you work at Vinted). Shifting the focus to returns on AI investment means concentrating more on outcomes-based measures, such as team productivity or customer satisfaction. But these require careful thought. Scientists, for example, are using AI to produce more papers (yay!) of dubious merit (boo!). A recent study by Tulsi Suchak of the University of Surrey and her co-authors found that papers derived from a health and nutrition data set ticked along at an average of four a year between 2014 and 2021; in the first nine months of 2024, there were 190. So outcomes-based measures have to be designed to capture both positive and negative effects of AI. At Google, for example, engineers have long been measured on three dimensions. Speed is one: how long does a code review take, say. Ease is another: how much friction is there in the system for things like onboarding new developers. The last, critically, is quality: “I don’t want to just go fast and make it easy to ship terrible software,” says Richard Seroter of Google Cloud. Complicating matters further is the fact that it often takes a while for the benefits of AI to show up. People have to spend time learning how to use the technology. Increased output can create bottlenecks elsewhere. A paper published last year by Kristina McElheran of the University of Toronto and her co-authors looked at AI adoption among American manufacturing firms, and confirmed the existence of a “J-curve” in which productivity dips before it leads to an improvement. This effect is especially pronounced at established firms and seems to be partly explained by changes to management practices that had previously helped firms to keep tabs on
performance. Any kind of calculation about projected returns has to take this initial period of disruption into account. Bosses also have to take a view on something much more fundamental: what kind of return are they interested in? One way to show financial returns on AI is to seize on time savings as an excuse to cut headcount. For some jobs and in some circumstances, that might make sense. But chainsaws and employee morale go together about as well as salt and slugs. And people who are no longer there cannot be redeployed in more productive ways. A better option is to calculate the amount of money that AI saves on expanding headcount. (Working out how much extra revenue can be attributed to the technology is even more of an art; at the very least, you need good baseline data and a culture of A/B testing to isolate its effects.) On top of everything, there is wild uncertainty about where AI is heading. That is an argument for a third set of measures, to go alongside ones on inputs and outcomes. These measures, which Mr Seroter calls “organisation- based”, are more geared towards capturing how well change is being managed. They could include data on levels of employee satisfaction with the way AI has been implemented, or progress towards building in-house expertise in areas that AI will not automate. A one-eyed focus on usage ignores the importance of returns. A rigid focus on returns ignores the need to keep learning. ■ Step inside the world of work with our Bartleby newsletter. Each week our white-collar oracle muses on the agonies of office life. This article was downloaded by zlibrary from https://www.economist.com/business/2026/08/20/how-to-measure-returns-on-ai
Meta’s blockbuster trial draws parallels to big tobacco But investors aren’t panicking yet Aug 20th 2026 IN EARLY October a sequel to “The Social Network”, a film about Mark Zuckerberg’s founding of Facebook, is scheduled for release. Called “The Social Reckoning”, it will dramatise the role of whistleblowers who alleged that the tech giant hid the fact that its products were harmful. Some of the characters involved will take part in a no less vivid drama involving Facebook, Instagram and their parent company, Meta, in a federal courtroom in Oakland, California. Oral arguments started on August 18th. Reckoning or not, the trial may be a blockbuster. California, Colorado, Kentucky and New Jersey are leading a group of 29 states that are suing Meta over allegations that its platforms developed
features harmful to children, and that it violated their privacy. In opening arguments, Megan O’Neill, a deputy attorney-general for California, said that Meta’s business model was to “hook the users, hold them for as long as they can, harvest their data and then hide the truth from the public.” Rob Bonta, California’s attorney-general, has described the case as Meta’s “tobacco moment”, likening the firm’s alleged deceptions to those that led to huge payouts by tobacco firms in the 1990s. Paul Schmidt, a lawyer for Meta, told jurors that the company took its responsibility to young users seriously, adding that there was no clear link between youngsters’ social-media use and ill health. Meta is expected to fight efforts by the plaintiffs to turn the trial into a referendum on the impact of social media on children, focusing instead on the narrow legal issues in dispute. Its lawyers have asked why it is being singled out, when youngsters’ use of YouTube, owned by Google, and TikTok, created by China’s ByteDance, is even higher than it is of Instagram. A lot is at stake. In a similar case brought by the government of New Mexico, a judge this month ordered Meta to pay almost $950m in penalties. The four states acting as plaintiffs in Oakland appear to be seeking more than 200 times as much if Meta loses the case, putting their estimate at $193bn—roughly equivalent to the firm’s revenue in 2025. If that were a benchmark for other states, it would devastate Meta. Whether it comes to that is another matter, though. In emotive language, Ms O’Neill alleged that Meta particularly targeted young children. She frequently used the word “hooked”. But unlike a trial in Los Angeles in March, in which Meta and Google, were obliged to pay damages to a 20- year-old harmed by spending much of her life on social media, the Oakland case is not focused on addiction. Instead it is what one of the attorneys-general calls “the largest consumer- protection lawsuit in American history”. Vincent Joralemon, of the University of California’s Berkeley Centre for Law & Technology, says the big question is whether or not Meta deceived the public about the impact of its products. That is similar to the tobacco cases. It also means whistleblower testimony is likely to be more relevant than thorny psychological debates about addiction, reckons Mr Joralemon.
Meta’s investors tend to think that the burden of proof on the deception charges is “quite high”, says Gil Luria of D.A. Davidson, an investment bank. Litigation risk has weighed on Meta’s valuation, contributing to an earnings ratio that is lower than those of its big-tech peers. But rather than fearing a giant payout, shareholders worry more about remedial changes that the prosecutors may seek if Meta loses in court, according to Mr Luria. Those include changes to features Meta, Google and TikTok use to keep user engagement high, such as “infinite scroll”, which continuously loads new posts. “If we didn’t have infinite scroll, we wouldn’t have as many ads,” Mr Luria says. Even if Meta loses, it has scope for appeal. It could, for instance, do so on the grounds that it is protected by Section 230 of the Communications Act, which safeguards internet platforms from legal liability for content posted by third parties. It sought to use that argument to halt the Oakland trial before it started, but the judge ruled that the appeal was premature. Public concern in America about the impact of social media on children is high, and being dragged through the courts may not help Meta’s image with parents. In order to head off similar worries about AI, on August 18th OpenAI launched ChatGPT for Teens, which strengthens guardrails for under 18-year-olds. For all the attention the Meta trial will receive, court verdicts can have less of an impact than the headlines suggest. The tobacco industry agreed in a settlement of 1998 to shell out more than $200bn over 25 years. But such outcomes have a “disappointing history” in America, says Matthew Lawrence of the Emory University School of Law. “We can mitigate the harms of industries exploiting addiction, but we have not yet had great success in finding ways to eliminate the harms.” ■ This article was downloaded by zlibrary from https://www.economist.com/business/2026/08/18/metas-blockbuster-trial-draws- parallels-to-big-tobacco
Can Reddit survive in the AI era? It wants more cash for its data—and more of people’s time Aug 20th 2026 Earlier this summer “Backrooms”, a horror flick, became one of the most successful independent films ever made, grossing $393m from a $10m budget. Its source material was unusual: a collection of images of unnervingly empty rooms that spread on Reddit, an internet forum. Hollywood producers have since taken to trawling the site for inspiration. Reddit has had a difficult time of late. Its share price has dropped by about a third this year. Strong quarterly earnings last month failed to cheer shareholders, who fret that the company will suffer as users turn to artificial- intelligence chatbots for information, rather than search engines such as Google that direct them to other sites. For now Reddit is attempting to cajole chatbot-makers into paying more for its data, both through deals and