The Future of Unpaid Work: AI's potential to transform unpaid domestic work in the UK and Japan

“工作的未来”近年来引起了人们的广泛关注。在了解新技术对未来就业和培训的影响方面作出了重大努力。相比之下，没有报酬的家务劳动的未来却很少受到关注。尽管英国工作年龄的成年人将56%的有偿和无偿工作时间花在家务和护理工作上；日本的这个数字是38%。人工智能技术取代家务劳动，可能会为所有年龄段的男性和女性腾出大量时间。它也有可能使花费在难以自动化的家庭工作上的时间更有效率，或者人们想自己做的事情（例如交互式儿童保育或老人护理），并有助于将其整合到更少的剧集中。该项目将把无偿家务劳动纳入人工智能和劳动力未来的讨论中，并评估其在两个社会和文化截然不同的国家的影响。为此，我们将在三个不同的工作包（WP）中考虑影响人工智能驱动的节省时间的国内技术扩散的多种因素。在WP1中，我们将首先使用2014-15年英国时间使用调查中确定的约50个此类任务的网格来评估家庭工作任务自动化的技术可能性，我们将应用广泛使用的弗雷-奥斯本方法的修改版本来测量自动化。然后，我们将使用一个专家小组来评估人工智能驱动的家庭技术将以多快的速度变得不仅在技术上可行，而且对家庭来说也是负担得起的。在WP2中，我们将解决影响家庭采用智能机器的社会因素。我们将首先分析购物的案例，通常在人工智能驱动的应用程序的帮助下进行离线或在线购物。英国和日本是电子商务的全球领导者，在这些国家进行网上购物是一个很好的测试案例，可以分析使用人工智能技术执行家庭任务与不使用人工智能技术执行家务有何不同。人工智能技术如何影响家庭不同成员的日常时间使用和活动参与率？然后，我们将进行一项实验性小调查，以评估将一系列国内任务外包给人工智能技术的可接受性。我们的小插曲将描述一个虚构的家庭情况，在这个家庭中执行一项给定的家务任务。在小插曲中，我们将随机改变一些核心因素，如任务类型，家庭收入，谁执行任务等。这些小插曲将辅以调查问题，询问当家庭成员不同意时如何决定采用国产技术。WP3涉及使用国家时间转移账户（NTTAs）中使用的方法模拟不同年龄的男性和女性无薪工作时间的可能变化。NTTAs使用时间使用数据来显示不同年龄组和性别的人如何在家庭中生产和消费用于各种无偿工作的时间，并在全国范围内估计这些群体之间的净时间转移。我们将在考虑到预期的人口变化和自动化的情况下，评估未来英国和日本国内任务自动化的情景，并评估未来跨性别和代际时间转移。我们将依赖于在WP1和WP2中开发的自动化、可负担性和可接受性分数。我们还将开发一个场景，在这个场景中，家庭任务的自动化将导致类似于WP2中分析的人工智能购物技术带来的变化。我们的分析将把目前主要由女性从事的无形家务劳动纳入“工作的未来”辩论。它将丰富我们对我们所拥有的最稀缺资源——时间——将如何受到人工智能技术的影响的理解。

The "Future of Work" has attracted much attention in recent years. Significant efforts are directed at understanding the implications of new technologies for the future of employment and training. In contrast, the future of unpaid domestic work has received little attention. This is although working-age adults in the UK spend about 56% of all paid and unpaid work time on household and care work; the figure for Japan is 38%. AI-powered technology replacing human labour in domestic tasks can potentially free up large amounts of time for men and women of all ages. It also has the potential to make the time spent on the domestic work that is hard to automate, or that people want to do themselves (e.g. interactive childcare or eldercare), more efficient and help consolidate it into fewer episodes. This project will bring unpaid domestic work into the discussion of AI and the future of labour and assess its implications in two socially and culturally quite distinct countries. To do this we will consider multiple factors that will influence the diffusion of AI-powered time-saving domestic technology in three distinct Work Packages (WP). In WP1 we will first evaluate the technological likelihood of automatibility of domestic work tasks using a grid of around 50 such tasks identified in the UK Time Use Survey 2014-15, to which we will apply a modified version of the widely used Frey-Osborne approach to measuring automatibility. We will then use an expert panel to assess how quickly AI-powered domestic technologies will become not only technologically possible, but also affordable for households. In WP2 we will address social factors affecting the adoption of intelligent machines at home. We will first analyse the case of shopping, commonly carried out offline or online with the help of AI-powered apps. UK and Japan are global leaders in e-commerce, making online shopping in these countries a good test case to analyse how performing a household task with AI-powered technology may differ from performing it without it. How does AI-powered technology affect daily time use and the rates of participation in the activity by different members of the household? We will then carry out an experimental vignette survey to evaluate the acceptability of outsourcing a range of domestic tasks to AI-powered technology. Our vignettes will describe a fictitious family situation in which a given domestic task is performed. Across the vignettes, we will randomly vary a number of core factors such as type of task, family income, who performs the task etc. The vignettes will be supplemented by survey questions asking how decisions to adopt domestic technology are made when family members disagree about it.WP3 involves simulations of likely changes in unpaid work time for men and women of different ages using the methodology used in National Time Transfer Accounts (NTTAs). NTTAs use time use data to show how people of different age groups and gender produce and consume time spent for various kinds of unpaid work in households, estimating net time transfers between such groups on the national level. We will estimate future scenarios of domestic task automation and evaluate future time transfers across genders and between generations in UK and Japan factoring in expected population change and automation. We will rely on automatibility, affordability and acceptability scores developed in WP1 and WP2. We will also develop a scenario in which automation of domestic tasks will result in changes similar to those brought about by AI-powered shopping technology as analysed in WP2. Our analysis will bring invisible domestic labour that today is largely performed by women into the "Future of Work" debate. It will enrich our understanding of how time, the scarcest resource we have, will be influenced by AI-powered technologies at home.