US Firms Punish Workers with Software 16x More Than Europe: OECD

A recent report from the Organisation for Economic Co-operation and Development (OECD) reveals a stark disparity in how employers in the United States and Europe utilize algorithmic management software, particularly for punitive measures. US firms are sixteen times more likely than their European counterparts to use software to sanction poor worker performance, a gap the OECD explicitly attributes to differing regulatory environments.

The groundbreaking survey, which polled 6,047 firms across six countries between June and August 2024, found that while the overall adoption of software for worker instruction, monitoring, and evaluation is high globally, its application varies dramatically. Ninety percent of US firms leverage these tools, contrasting with an average of 78% in the surveyed European countries (France, Germany, Italy, and Spain) and 40% in Japan.

A Chasm in Algorithmic Control

The most significant finding is the "evaluation gap." Software designed to sanction poor performance is employed by a staggering 67% of US companies, contrasting sharply with just 4% in Europe and 1% in Japan. Similarly, software for rewarding good performance is used by 83% of US firms versus 13% in Europe. Performance leaderboards, visible to staff, are present in half of US workplaces but only 7% of European ones.

Beyond evaluation, monitoring practices also show a wide divide. Seventy-two percent of US firms track the speed of work, compared to 15% in Europe. Even more invasively, 55% of US companies monitor the content and tone of employee conversations, calls, or emails, a practice seen in only 6% of European firms.

This isn't merely about presence but intensity. More than three-quarters of American firms utilize ten or more distinct algorithmic management tools, with very few using fewer than eight. European companies, in contrast, typically employ only three to five, while nearly a third of Japanese adopters use just one.

Regulation Drives the Divide

The OECD points directly to regulatory architecture as the cause of this transatlantic difference. Europe's centralized, rights-based approach, underpinned by directives requiring information and consultation with worker representatives, stands in stark contrast to the US's fragmented system of agency enforcement and state and local regulations. In the US, bargaining obligations are often limited to specific subjects, and while managers claim employee consultation, actual worker awareness is likely much lower.

Intriguingly, within Europe, Italy and Spain demonstrate superior governance, with 83% and 78% of their firms conducting audits of these tools, significantly higher than France and Germany. This success is credited to targeted national laws, such as Italy's 2022 Transparency Decree covering automated decision-making and Spain's 2021 Riders’ Law, which grants works councils the right to information on algorithmic management. Robust enforcement, exemplified by the €2.6 million fine against Glovo’s Italian subsidiary (Foodinho) under data protection law, has further bolstered these measures.

The AI Act's Blind Spot

A critical warning from the OECD is that the vast majority of these algorithmic management tools are not powered by artificial intelligence. Instead, they are often standard enterprise software like SAP, Workday, Oracle, Jira, Asana, and time-tracking systems that have quietly acquired evaluation or monitoring functions. This poses a significant challenge for the upcoming EU AI Act, which is specifically designed for AI systems.

Rules drafted with AI in mind risk overlooking widespread practices facilitated by conventional software, creating a regulatory loophole. Employers often don't even recognize or brand these integrated functions as 'algorithmic management,' further complicating efforts to regulate them effectively.

Concerns about worker awareness are also high. Despite 91% of managers claiming employees are informed, the OECD expresses doubt, suggesting a disconnect between management perception and worker reality. Moreover, 69% of European managers anomalously stated their tools don't collect individual worker data, which the OECD attributes to a lack of understanding or careful GDPR-influenced responses. In the US, 90% of managers indicate workers cannot opt out of data collection, and over half report workers cannot request corrections.

Looking Ahead

While the survey has limitations, notably the absence of worker perspectives and varying response rates (e.g., 2% in Germany, 6% in Japan), its findings underscore an urgent need for policy intervention. The debate on algorithmic management has been reignited within the European Commission, with calls for extending the Platform Work Directive beyond gig-economy platforms. Future regulations must develop robust transparency duties, ensuring they have practical enforcement mechanisms that extend beyond AI-specific definitions to cover all forms of algorithmic oversight in the workplace.

FAQ

Q: What is algorithmic management in the workplace?

A: Algorithmic management refers to the use of software and algorithms to instruct, monitor, evaluate, and even sanction employees. This can include tools for scheduling, tracking work speed, monitoring communications, and assessing performance.

Q: Why is there such a large difference in software-based punishment rates between the US and Europe?

A: The OECD attributes this gap primarily to regulatory architecture. European countries generally have more centralized, rights-based labor laws that mandate worker consultation and offer stronger data protection, whereas the US has a more fragmented regulatory landscape.

Q: Does the EU's new AI Act address these issues related to algorithmic management?

A: The OECD warns that most algorithmic management tools currently in use are not AI-powered. Therefore, the EU AI Act, which specifically targets AI systems, may not cover the majority of software used by employers for monitoring and sanctioning workers, creating a potential regulatory blind spot.