Human listeners and virtual assistants: privacy and labor arbitrage in the production of smart technologies

I’m glad to announce the publication of new research, as a chapter in the fabulous Digital Work in the Planetary Market, a volume edited by Mark Graham and Fabian Ferrari and published in open access by MIT Press.

The chapter, co-authored with Antonio A. Casilli, starts by recalling how In spring 2019, public outcry followed media revelations that major producers of voice assistants recruit human operators to transcribe and label users’ conversations. These high-profile cases uncovered the paradoxically labor-intensive nature of automation, ultimate cause of the highly criticized privacy violations.

The development of smart solutions requires large amounts of human work. Sub-contracted on demand through digital platforms and usually paid by piecework, myriad online “micro-workers” annotate, tag, and sort the data used to prepare and calibrate algorithms. These humans are also needed to check outputs – such as automated transcriptions of users’ conversations with their virtual assistant – and to make corrections if needed, sometimes in real time. The data that they process include personal information, of which voice is an example.

We show that the platform system exposes both consumers and micro-workers to high risks. Because producers of smart devices conceal the role of humans behind automation, users underestimate the degree to which their privacy is challenged. As a result, they might unwittingly let their virtual assistant capture children’s voices, friends’ names and addresses, or details of their intimate life. Conversely, the micro-workers who hear or transcribe this information face the moral challenge of taking the role of intruders, and bear the burden of maintaining confidentiality. Through outsourcing, platforms often leave them without sufficient safeguards and guidelines, and may even shift onto them the responsibility to protect the personal data they happen to handle.

Besides, micro-workers themselves release their personal data to platforms. The tasks they do include, for example, recording utterances for the needs of virtual assistants that need large sets of, say, ways to ask about the weather to “learn” to recognize such requests. Workers’ voices, identities and profiles are personal data that clients and platforms collect, store and re-use. With many actors in the loop, privacy safeguards are looser and transparency is harder to ensure. Lack of visibility, not to mention of collective organization, prevents workers from taking action.

Note: Description of one labor-intensive data supply chain. A producer of smart speakers located in the US outsources AI verification to a Chinese platform (1) that relies on a Japanese online service (2) and a Spanish sub-contractor (3) to recruit workers in France (4). Workers are supervised by an Italian company (5), and sign up to a microtask platform managed by the lead firm in the US (6). Source: Authors’ elaboration.

These issues become more severe when micro-tasks are subcontracted to countries where labor costs are low. Globalization enables international platforms to allocate tasks for European and North American clients to workers in Southeast Asia, Africa, and Latin America. This global labor arbitrage goes hand in hand with a global privacy one, as data are channeled to countries where privacy and data protection laws provide uneven levels of protection. Thus, we conclude that any solution must be dual – protecting workers to protect users.

The chapter is available in open access here.

Unboxing AI conference

I’m excited to be part of the organizing team for an upcoming conference entitled “Unboxing AI” and aiming to open – at least to an extent – the black box. What are the material conditions of AI production? Who are the multitudes of precarious workers who contribute to it in the shadow, by generating data and checking algorithmic outputs? What are the geographical areas and the social scope of the work that produces today’s intelligent technologies? These are some of the questions we aim to explore.

The first two days of the conference (November 5 and 6, 3pm – 7pm CET) will bring together highly regarded international specialists from a wide variety of disciplines (sociology, law, economics, but also the arts and humanities…). On the third day (November 7, also 3 pm – 7 pm CET), there will be a doctoral colloquium with a selection of very promising work by young researchers.

The conference was initially planned to take place in Milan in March 2020, and had to be postponed due to the Covid-19 pandemic. As the health situation is still critical, we have opted for an online-only version. At least, this format is cheap – no need to travel to attend – and we can welcome a more geographically diverse range of participants. Indeed the afternoon-only schedule is meant to enable colleagues from North and South America to attend.

Participation is free of charge but prior registration is required. You will find the programme as well as registration forms here
(please note that there is a separate form for each of the three dates of the conference).

The conference is organized as part of the initiatives of our ‘International Network on Digital Labor‘ and is co-sponsored by ISRF (Independent Social Research Foundation), the Nexa Center for Internet and Society, and Fondazione Feltrinelli.

Crowdworking Symposium 2020

With Antonio A. Casilli, I will be presenting a paper tomorrow at the Crowdworking Symposium organized by the University of Paderborn, Germany. Unfortunately, we will participate only online because of the health situation.

Our mini-paper (3 pages), entitled ‘Portraits of micro-workers: The real people behind AI in France’, is available here.

The platform economy, labour and Covid-19

On 18 September 2020, I present my research on the platform economy and its impact on labour in Covid-19 times at Nantes Digital Week, as part of a special event organized by CGT, a Union.

The mobility restrictions that accompanied the pandemic encouraged use of digital tools to socialize, study and work, suggesting that automation is gaining ground and that technology enables contactless – hence safe – interactions in much of our social life. Yet behind apparent automation, precarious and unprotected human labour is hidden. Workers recruited through digital platforms to make these solutions work, are in fact disproportionately exposed to risks. I illustrate these ideas in three main cases: food delivery workers, that enabled the restaurant industry to stand the crisis even during lockdown; commercial content moderators that are to return to office sooner than others, to protect our safety online; and AI micro-workers who trained tools whose sales have gone up during stay-at-home rules, such as voice assistants, and helped the creation of datasets for much-needed health applications.

PhD studentship available

I have an exciting opportunity for a brilliant master’s degree holder willing to do a PhD in economic sociology. The topic of the thesis is “The division of data labour: How multi-level micro-work networks elucidate the social and economic dimensions of artificial intelligence”. The studentship is generously funded by CNRS.

More information and application form (both in French and in English) here: https://emploi.cnrs.fr/Offres/Doctorant/UMR8623-PAOTUB-001/Default.aspx?lang=EN

The trainer, the verifier, the imitator: Three ways in which human platform workers support artificial intelligence

New article, co-authored with Antonio A. Casilli and Marion Coville, just published in Big Data & Society!

The paper sheds light on the role of digital platform labour in the development of today’s artificial intelligence, predicated on data-intensive machine learning algorithms. We uncover the specific ways in which outsourcing of data tasks to myriad ‘micro-workers’, recruited and managed through specialized platforms, powers virtual assistants, self-driving vehicles and connected objects. Using qualitative data from multiple sources, we show that micro-work performs a variety of functions, between three poles that we label, respectively, ‘artificial intelligence preparation’, ‘artificial intelligence verification’ and ‘artificial intelligence impersonation’. Because of the wide scope of application of micro-work, it is a structural component of contemporary artificial intelligence production processes – not an ephemeral form of support that may vanish once the technology reaches maturity stage. Through the lens of micro-work, we prefigure the policy implications of a future in which data technologies do not replace human workforce but imply its marginalization and precariousness.

The three main functions of micro-work in the development of data-intensive, machine-learning based AI solutions.

The paper reports results of the 2017-18 DiPLab project, and is available here in open access.

Internship offer, TRIA project

I am currently seeking to hire a student intern for new research project TRIA (Les TRavailleurs de l’Intelligence Artificielle / Los TRabajadores de la Inteligencia Artificial). Start as soon as possible, conditional on evolving regulations at the end of the current lockdown. Max 6 months.

A full description of the project is enclosed (in French).

Microwork platforms: a challenge for artificial intelligence, a challenge for employment?

With our sponsors France Stratégie and MSH Paris-Saclay, we convene an international conference on micro-work in Paris on June 13, 2019, followed by the first INDL (International Network on Digital Labor) workshop on June 14. The event will include a “meet the microworkers” panel on June 13, where workers, platform owners and client companies will take the stage. There will also be presentations of the results of national and international surveys (notably ours, DiPLab) on these emerging forms of work, and discussions with French and international academic and institutional experts.

Bandeau-SITE-Microtravail-FR-1280x640

After Uber, Deliveroo and other on-demand services, micro-work is a new form of labor mediated by digital platforms. Internet and mobile services recruit crowds to perform small, standardized and repetitive tasks on behalf of corporate clients, in return for fees ranging from few cents to few euros. These tasks generally require low skills: taking a picture in a store, recognizing and classifying images, transcribing bits of text, formatting an electronic file… Despite their apparent simplicity, these micro-tasks performed by millions of people around the world, are crucial to create the databases needed to calibrate and “train” artificial intelligence algorithms.

Internationally, Amazon Mechanical Turk is the most widely known micro-work platform. In France and in French-speaking Africa, other platforms are attracting a growing number of workers to supplement or even substitute for their primary income. How widespread is the phenomenon? How to recognize, organize and regulate this new form of work? How, finally, does it relate to traditional forms of employment?

Presentations and discussions are held in French and English, with simultaneous translation.

The programme is available here.

Back from Reshaping Work 2018

I was last week at the second Reshaping Work in the Platform Economy in Amsterdam. The interest of this small conference is tht it brings together different actors of the platform economy, from academics and students to policymakers, union leaders,  workers, and representatives of platforms to discuss.

In an overview of preliminary results of our project DiPLab, Antonio A. Casilli and I presented our reflection on how micro-work powers artificial intelligence (AI), in three main ways:

  1. Training AI
  2. Validating outcomes of AI
  3. Impersonating AI when it is cheaper or simpler that real AI

AI

No more details for now… it will come out as a working paper very soon!