Group items tagged

People Analytics avait tenté d'automatiser en partie ce processus, en élaborant un algorithme (ci-contre) qui se reposait sur les grilles d'évaluation et leur attribuait un coefficient. Il a alors découvert qu'il pouvait automatiser les promotions sur près d'un tiers des dossiers, et pensait que cette nouvelle réjouirait les ingénieurs qui sont habitués à manipuler chaque jour des algorithmes, quasi religieusement. Mais non. La proposition de People Analytics s'est heurtée à une résistance des cadres, qui ont refusé de traiter le sort des hommes et des femmes par une IA. Tout au plus ont-ils reconnu que l'algorithme pouvait les aider à identifier d'éventuels dossiers qu'ils n'auraient pas bien examiné.

y 2030, up to 30% of the hours worked globally could be automated, according to a new report by the McKinsey Global Institute. Analysts in the consultancy’s research arm estimate that between 400 million and 800 million people could find themselves displaced by automation and in need of new jobs, depending on how quickly new technologies are adopted. Of this group, as many as 375 million people—about 14% of the global workforce—may need to completely switch occupational categories and learn a new set of skills to find work.

Notably, McKinsey argues that demand for work will increase as automation grows. Technology will drive productivity growth, which will in turn lead to rising incomes and consumption, especially in developing countries. Meanwhile, there will be more jobs in health care to meet the demands of aging societies and more investment in infrastructure and energy.

For these benefits to be realised, everyone needs to gain new skills, with governments and private companies taking on the unprecedented task of retraining millions of people in the middle of their careers. “Even if there is enough work to ensure full employment by 2030, major transitions lie ahead that could match or even exceed the scale of historical shifts out of agriculture and manufacturing,” the report says.

The line between automation and humans blurs more often than Silicon Valley might like to admit. Facebook hired thousands of people this year to moderate content on its social network, after algorithms repeatedly failed to do the job. Uber depends on more than 2 million drivers worldwide to provide rides every day, as well as employees at headquarters to make sure enough of those drivers are on the road. Behind much of Google’s digitization of books and maps is random people on the internet, conscripted using reCaptcha. Expensify is just another example.

The receipts on Mechanical Turk belonged to “less than 0.00004% of users—none of whom are paying customers,” Barrett said, adding that, at any rate, there is nothing important on a receipt, “that’s why receipts are so commonly thrown out—because they are literally garbage.” Also: “anybody concerned by the real-world risks of a vetted, tested transcriptionist reading their Uber receipt should probably consider the vastly more immediate and life-threatening consequences of getting into that stranger’s car in the first place.”