The tyranny of targets - FT中文网
登录×
电子邮件/用户名
密码
记住我
请输入邮箱和密码进行绑定操作:
请输入手机号码,通过短信验证(目前仅支持中国大陆地区的手机号):
请您阅读我们的用户注册协议隐私权保护政策,点击下方按钮即视为您接受。
卧底经济学家

The tyranny of targets

Not only is quantification changing our behaviour, but we’re failing to notice this has happened
00:00

{"text":[[{"start":4.73,"text":"I recently described the contradictions inherent in my fitness-tracking watch. On the one hand, it had unlocked the joy of running for me, encouraging me to run further and faster and set goals I’d never dreamt of achieving. On the other, the watch could also push me into counter-productive behaviour, such as running through injury — and had a tendency to turn a pleasant run into a quantified grind."}],[{"start":32.519999999999996,"text":"What an eye-opening delight, then, to pick up C Thi Nguyen’s book The Score. Nguyen is a philosopher and, more importantly, an enthusiast for all sorts of playful activity from rock-climbing to yo-yo tricks to role-playing games. He argues that the contradiction I found in my fitness watch is part of a much bigger story about the way we have allowed quantified metrics to seep into so many parts of life."}],[{"start":60.47,"text":"The tyranny of targets is familiar stuff to social scientists, even if it seems to be a source of endless unpleasant surprises for policymakers and corporate executives. Charles Goodhart, an economist, and Donald Campbell, a psychologist, each famously explained that perfectly decent quantitative metrics become corrupted once they’re pressed into service as targets."}],[{"start":86.8,"text":"Nguyen agrees, but wants to go further. The problem is not just that we contort our behaviour to hit the target in the service of some bureaucratic incentive; it’s that we cease to realise that we are contorting our behaviour at all. The metrics, which are inevitably reductive, begin to seep into our own judgment of what matters. Nguyen calls this value capture: we abandon our rich and subtle personal values in favour of the simplified, quantified values we are being so conveniently offered."}],[{"start":120.13,"text":"Why does this happen? A simple answer is “because of computers”, but there is more to it than that. The rise of quantification isn’t the result of some tech-bro conspiracy. It’s been happening for a long time and for some good reasons."}],[{"start":136.45999999999998,"text":"First, quantified evaluations — of a student essay, a surgeon’s safety record or a bottle of wine — can be added up, averaged, analysed and compared across contexts. They become portable: you can compare one school class to another, one hospital to another, one wine producer to another. Everything can be analysed from a distance."}],[{"start":161.79999999999998,"text":"Second, quantification is often the result of algorithmic rules. Take the inputs, crank the algorithmic handle and out come the outputs. If you’re a university president wanting your institution to rise up the rankings, there are levers to pull: persuade more candidates to apply so that you can be more selective, spend money on the library and help your graduates find jobs. These rules are accessible: you don’t need deep expertise to apply them or to interpret them."}],[{"start":195.98999999999998,"text":"Third, these algorithmic procedures tend to make people interchangeable. Rather than relying on the ineffable intuitions of a teacher or a doctor or a coach, you have an objective, consistent process: any qualified person can follow the quantified process and so all qualified people are substitutes for each other."}],[{"start":219.42,"text":"Fourth, all these quantified metrics allow a centralised authority to co-ordinate and control what would otherwise be a mess of idiosyncratic decisions."}],[{"start":230.97,"text":"Portability, accessibility, interchangeability and co-ordination are often huge advantages. It is no exaggeration to suggest that modern life, including modern science, modern technology, modern politics and the modern economy, depends on these four key features of quantification."}],[{"start":253.8,"text":"But each of the four has a downside. While quantification allows us to make comparisons across different contexts, it also strips away those contexts. Rules-based decisions are intelligible and accessible to non-experts, but they are also rigid, failing to adapt to the subtleties of circumstance. Making people interchangeable does away with unwelcome inconsistencies, but also robs decision-making of individual judgment and sensitivity. And the downside of co-ordination is the loss of autonomy."}],[{"start":288.18,"text":"There are good metrics and bad ones, of course. Some are corrupt, counter-productive and riddled with perverse incentives. Others impose method on madness, identifying more effective medicines, more productive manufacturing techniques or more profitable trades. "}],[{"start":null,"text":"


"}],[{"start":308.22,"text":"Still, even the best metrics are inevitably reductive, stripping away context, discretion and judgment. These metrics also strip away diversity, forcing everyone to agree on what is best, even when we should let a thousand flowers bloom. In the pre-metric era, a university might have had a distinctive outlook, aiming to serve a particular kind of student in a particular way. Once each university wants to rise up the rankings they will need to embrace whatever values the rankings enshrine. If the metrics are good ones, then quality might rise, but the range of offerings will shrink. If the metrics are bad ones, then both quality and diversity will suffer."}],[{"start":354.67,"text":"There may be little point in trying to turn the tide of quantification in business and policy, but we can at least try to prevent it seeping into our souls. Yet why is it so easy to let the data rule the way we sleep, the way we move, the way we eat and even the way we interact our friends? Yes, digital devices have been designed to twiddle our dials. But the world is also a confusing place and we are hungry for simple, accessible guidance. Rely too much on the metrics, however, and we risk outsourcing our judgment and replacing it with thin, globalised values that do not really reflect who we are."}],[{"start":395.71000000000004,"text":"Like Nguyen, I love games, and games have rules and a score. Why do they feel so different from the gamified world in which we chase streaks on our language-learning apps, try to move the VO2 needle on our fitness watches and, at work, crank away at some corporate performance metric with one eye on promotion?"}],[{"start":416.50000000000006,"text":"Perhaps the secret to a good game or sport is that the goals are both tantalisingly challenging and utterly unimportant: that’s true of my marathon, Nguyen’s rock-climbing, a game of football or a game of chess. What makes the game fun is that it’s so hard to succeed — and that whether you succeed or fail, you can always play another game."}],[{"start":441.84000000000003,"text":"In a fun game, the rules are a scaffolding for playful mastery. And in a gamified life? The rules may be a scaffolding for productivity — but all too often, they are a cage for us poor players."}],[{"start":457.07000000000005,"text":"Find out about our latest stories first — follow FT Weekend Magazine on X and FT Weekend on Instagram"}],[{"start":472.6700000000001,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1770962957_3683.mp3"}

版权声明:本文版权归FT中文网所有,未经允许任何单位或个人不得转载,复制或以任何其他方式使用本文全部或部分,侵权必究。

AI会加剧劳动者之间的不平等吗?

这些以及更多发现,均来自我们对数千名美英员工人工智能使用情况的新调查。

这届世界杯错失了一个“空门”

我对此最有发言权——身为墨西哥驻美大使,我曾参与筹划最初的申办。

如果AI无法自我辩护,它还能进行歧视吗?

埃隆•马斯克对科罗拉多提起的诉讼,揭示了关于人工智能与民主更深层的哲学问题。

工党究竟取得了什么成就?

基尔•斯塔默已将其多数优势转化为数项重要法律,但公众几乎无感并非偶然。

小米高端电动车进军欧洲,挑战特斯拉

在对其汽车的需求超过产能之际,小米希望在欧洲大陆扩大版图。

谷歌押注AI优势,追赶云端对手亚马逊与微软

谷歌云首席执行官托马斯•库里安表示,其AI芯片和模型能够助力数据中心业务取得进展。
设置字号×
最小
较小
默认
较大
最大
分享×