Bridging the Workers' Data Value Gap in the Age of Corporate Automation

Abstract

The rapid advancement of artificial intelligence ("Al") is transforming corporate innovation, enabling the automation of work traditionally performed by humans.

Companies increasingly rely on internal data to develop and train Al systems, much of which originates from their own workforces. This includes work products created by employees or contractors in the course of their duties, here referred to as

"workers' data." Workers' data may hold significant value as it is often of high quality, high quantity and of high relevance. Yet, ironically, if workers' data is used for automation purposes, it could displace the very employees who were responsible for generating the data in the first place. The risk of a "job apocalypse," where hundreds of millions of jobs could be replaced by Al in the coming years, becomes increasingly real as more companies push towards automating parts of their workforces.

Work products used to generate workers' data will often qualify as copyright-protected works. The workers, who are authors in copyright law, will often have assigned their copyright to their respective employers or contractees, whether by statute or by contract. However, that companies may own the copyright for the work products, including the data, does not necessarily confer unlimited freedom of use. This Article critically examines the legal boundaries on the use of workers' data for Al and automation purposes, focusing on the European Union (EU), the United Kingdom (UK), and the United States. Overall, the current legal framework largely falls short of adequately protecting workers' rights when it comes to use of their data by employers, with one important exception. In the EU, the DSM Directive introduces a right for authors to claim additional remuneration where the economic value derived from their works is significantly greater than what they were originally paid. This right to contract adjustment, commonly referred to as the "best-seller" rule, may, in certain cases, help workers whose data are being used without fair remuneration, bridging what is coined as the "workers' data value gap." Whether workers are entitled to further remuneration will depend on highly fact-specific circumstances, and outcomes are likely to vary from case to case.

The urgency of addressing these issues cannot be overstated. Workers' data is expected to play an increasingly central role in corporate automation projects worldwide, with consequences that extend far beyond the EU. Without a clear and coherent legal framework, there is a dual risk. On the one hand, companies may undervalue or fail to properly compensate for the human contributions that make automation possible; on the other hand, legal uncertainty or excessive compensation claims could deter or slow down automation initiatives. This Article calls for economic, policy, and legal research to explore these questions in greater detail, and to develop balanced solutions that both protect workers' rights and foster innovation in the coming age of corporate automation.