Algorithmic management, software that automates or supports managerial tasks, is widely adopted across countries covered by a recent OECD employer survey. It is particularly prevalent in the United States, where 90% of managers say their firms have adopted at least one tool to instruct, monitor or evaluate workers. Adoption is slightly lower in the European countries surveyed (France, Germany, Italy and Spain, with average adoption of 79%) and moderate in Japan (40%).
In most countries, firms are less likely to adopt algorithmic management tools for monitoring worker activity that may collect personal data, such as health information, and tools for evaluating workers that are likely to affect more consequential outcomes, such as access to work. However, an exception is the United States, where algorithmic management tools of all types, including tools to monitor workers and evaluate workers, are highly prevalent.
Algorithmic management tools also appear to benefit how managers do their work. Managers using the tools perceive that their use improves the quality of their decision making. This is driven by increases in the information available to make decisions, increases in the speed of decision making, and increases in the autonomy to make decisions.
On the other hand, the introduction of algorithmic management tools is not without risk. Nearly two‑thirds of managers that use algorithmic management tools have concerns regarding their impact on workers. Across all countries, unclear accountability in the case of a wrong decision was the most reported issue, followed by inability to follow the logic of algorithmic decisions and inadequate protection of workers’ physical and mental health.
The introduction of algorithmic management tools should be guided by worker consultation to aid adoption and safeguard worker well-being. Moreover, in all country settings, it is important to examine and monitor current applicable legislation to assess its adequacy to protect against the potential harms of new technologies.
How widespread is algorithmic management in workplaces?
A recent OECD survey provided the first cross-country, representative data on the adoption of algorithmic management tools by firms based on an employer survey of over 6 000 mid-level managers across six countries: France, Germany, Italy, Japan, Spain and the United States. This policy brief summarises the key findings and highlights the importance of social dialogue around the introduction of new technologies in the workplace and the need to monitor the policy landscape for regulatory gaps.
Algorithmic management is the use of technological tools to fully or partially automate tasks traditionally carried out by human managers, including the collection of worker data without the processing of such data. Given the breadth of this definition, researchers often separate algorithmic management tools into sub-categories reflecting how they are used. The OECD survey gathered information on three sub-categories: (i) tools to instruct workers, (ii) tools to monitor workers, and (iii) tools to evaluate workers. Each sub-category contained specific use cases, as shown in Table 1, which were used to identify whether firms were adopters of algorithmic management tools.
Source: Milanez, Lemmens and Ruggiu (2025[1]), Algorithmic management in the workplace: New evidence from an OECD employer survey, https://doi.org/10.1787/287c13c4-en.
Algorithmic management tools range in their level of technological sophistication and are not necessarily AI-powered. This distinction is important as regulation covering the use of AI technologies, such as the EU AI Act, pertain to a subset of algorithmic management tools but not all.
Algorithmic management is widespread in most of the countries studied. In the United States, 90% of firms have adopted at least one tool to instruct, monitor or evaluate workers. The average adoption rate in the European countries surveyed was 79% (81% in France, 78% in Germany, 76% in Italy and 78% in Spain). This suggests that algorithmic management is more prevalent than previously understood and makes clear that managerial roles are evolving significantly through task automation or augmentation.1
In the United States, the widespread adoption of algorithmic management is intensified further by firms’ tendencies to have adopted many tools across each of the different categories. More than three‑quarters of US managers report that their firms provide ten or more of the 15 algorithmic management tools shown in Table 1, while very few report that their firms provide fewer than eight. European firms tend to adopt a more moderate number of algorithmic management tools compared to those in the United States.
The picture is different in Japan where, with a less digitalised business sector, only 40% of firms have adopted algorithmic management tools. Even when tools are adopted, adoption is less intense, with most Japanese firms using only one type of algorithmic management tool.
There also are distinct country patterns in the types of algorithmic management tools, as shown in Figure 1. In the United States, firms commonly use tools of all types, with average adoption rates of 90% across instruction, monitoring and evaluation tools. In the European countries surveyed, instruction tools were most prevalent overall (adoption rate of 69%, ranging from 64% in Italy to 74% in France), followed by monitoring tools (adoption rate of 67%, ranging from 64% in Germany to 71% in France) and lastly tools used to evaluate workers’ performance (adoption rate of 35%, ranging from 27% in Italy to 40% in France). In Japan, monitoring tools are slightly more common (adoption rate of 32%) than instruction and evaluation tools (25% and 11%, respectively).
Percentage of managers whose firms provide algorithmic management by tool sub-category and by country
Source: Milanez, Lemmens and Ruggiu (2025[1]), Algorithmic management in the workplace: New evidence from an OECD employer survey, https://doi.org/10.1787/287c13c4-en.
Among non-adopters, high cost was the leading reason for non-adoption (across all countries). In France, Germany, Italy, Japan, and Spain, staff resistance was the second-most cited reason, followed by concern for workers, which resonates with the findings above as well as with other research showing that workers can be detrimentally impacted by the use of algorithmic management tools. In the United States, the second-most cited reason for non-adoption was a lack of skills.
Broadly speaking, algorithmic management tools that may make greater use of personal data, such as health information, or that may impact more consequential outcomes, such as access to work or pay, have lower adoption rates. For example, in the category of monitoring tools, firms in European countries are less likely to adopt tools that involve the collection of data such as information derived from phone conversations or emails, health information or location, compared to tools that gather information on strictly work-related activity such as work time, completion of work activities or work speed. Within the category of evaluation tools, tools to sanction poor performance are less commonly adopted in every country compared to those to reward good performance (across all countries, adoption rate of 14% for sanctioning versus 23% for rewarding). Lower adoption rates for such algorithmic management tools can be due to a range of factors (e.g. regulation, firms’ and workers’ attitudes, social dialogue).
As algorithmic management tools automate or augment managerial tasks, changes in skills needs paint a picture of an evolved managerial role. Managers report the increased importance of analytical skills (60% report increased need, set against 3% who report decreased need), such as the ability to use or interpret data, problem solving skills, and digital skills. They also report increased importance of social skills (32% report increased need), including active listening skills, conflict resolution skills, empathy, and communication skills. Thus, managerial roles appear to require sufficient analytical skills to effectively utilise algorithmic tools but also require uniquely human qualities.
Managers have the perception that the use of algorithmic management improves the quality of their own decision making, with 60% citing improvement. This is driven by increases in the information available to managers to make decisions, increases in the speed of decision making, and increases in the autonomy of managers to make decisions, illuminating one channel whereby algorithmic management results in productivity and efficiency gains for firms. Beyond benefiting firms, improved managerial decision making may also benefit workers through greater consistency and objectivity of decisions.
However, managers have contrasting opinions in an area where algorithmic management has been touted for its promise: the potential to eliminate human bias from managerial decisions. While managers in European countries and Japan are more likely to believe that there is no effect or a net increase in bias, US managers commonly believe that algorithmic management tools decrease decision bias.
In contrast to the positive impact on decision making, nearly two‑thirds of managers that use algorithmic management tools have at least one concern regarding their use. Managers were asked whether they observed problems regarding different aspects of trustworthiness of algorithmic management tools in their workplaces, including bias, explainability, accountability, health and transparency. The most reported concern was unclear accountability in the case of a wrong decision (28% of managers raise this issue), followed by inability to follow the logic of algorithmic decisions or recommendations (27%) and inadequate protection of workers’ physical and mental health (27%) (Figure 2).
Percentage of managers reporting a specific trustworthiness concern
Note: Managers were asked whether they agreed or disagreed with statements on the safe and ethical use of software to automate the co‑ordination of work activities in their workplaces. The results correspond to the share of managers that answered they observed at least one of the following problems: bias and discrimination, lack of explainability, unclear accountability, inadequate protection of workers’ physical and mental health, workers are not made aware of the use of the software. Only users of algorithmic management are included in the results.
Source: Milanez, Lemmens and Ruggiu (2025[1]), Algorithmic management in the workplace: New evidence from an OECD employer survey, https://doi.org/10.1787/287c13c4-en.
Concerns appear to be linked to both how intensely algorithmic management tools are adopted and the types of tools used. For instance, the United States, with the highest overall adoption of algorithmic management tools and relatively high use of tools to monitor and evaluate workers, has the largest share of managers reporting trustworthiness concerns.
In the face of concerns regarding the trust use of algorithmic management, it is encouraging that the vast majority of firms attempt to govern such tools via a range of firm-level measures (impact or risk assessments, regular audits of software, guidelines for implementation or use, the existence of an ethics officer or ethics board, the existence of a whistleblower or complaint channel, or consultation of workers or their representatives). Nearly 90% of managers report that their firms have at least one measure in place. Guidelines were the most common measure (implemented by between 46% and 91% of firms), followed by worker consultation (almost two‑thirds of firms). However, more research is needed to understand the key components of various governance measures and to measure their efficacy to share best practices across firms and countries.
Worker consultation can play a key role in ensuring smoother transitions when new technologies are introduced into the workplace. By providing workers with the opportunity to voice concerns and give feedback, consultations can help mitigate risks and support flexible, pragmatic solutions to address adjustments in wages, staffing, work organisation, and training. Furthermore, worker consultation can boost engagement with and acceptance of new technologies.
Irrespective of the country setting, policymakers must first ensure that algorithmic management tools comply with existing legislation and standards. Countries vary widely in their approaches to regulating algorithmic management software, from the centralised, rights-based approach of the European Union to, in the United States, a patchwork of agency enforcement and state and local rules. Enforcement requires labour authorities and other authorities, such as data protection, to have adequate capacity to supervise, enforce compliance, and ensure that rules are not easily circumvented.
Next, there is a need for policymakers to identify and address gaps in legislation governing automated decision making. Debate around the potential need for new legislation is especially active in Europe following European Commission President von der Leyen’s call for an initiative on algorithmic management. Key areas being considered include transparency about the use of algorithmic management and right to information, protection of worker data, and potential prohibition of certain data collection (Barslund et al., 2025[1]).
These efforts are consistent with the OECD’s AI Principles (OECD, 2025[2]), the first intergovernmental standard on AI released in 2019 and updated in 2024. The Principles guide AI actors in their efforts to develop trustworthy AI and provide policymakers with recommendations for effective AI policies. The principles on transparency and explainability, robustness, security and safety, and accountability are particularly relevant to the development and implementation of algorithmic management software:
Transparency and explainability: Responsible disclosure around AI systems to ensure that people understand when they are engaging with them and can challenge outcomes;
Robustness, security and safety: AI systems must function in a robust, secure and safe way throughout their lifetimes, and potential risks should be continually assessed and managed; and
Accountability: Organisations and individuals developing, deploying or operating AI systems should be held accountable for their proper functioning in line with the OECD’s values-based principles for AI.
Crucially, relying on shared international standards and definitions promotes policy coherence across jurisdictions, reduces uncertainty for firms operating globally, and supports evidence‑based regulation. In the rapidly evolving domain of algorithmic management – where terms, practices, and risks can vary widely – these common reference points help ensure that new policy proposals are aligned with broader global norms and grounded in a consistent understanding of what responsible AI deployment should entail.
Milanez, A., A. Lemmens and C. Ruggiu (2025), “Algorithmic management in the workplace: New evidence from an OECD employer survey”, OECD Artificial Intelligence Papers, No. 31, OECD Publishing, Paris, https://doi.org/10.1787/287c13c4-en.
References
[1] Barslund, M. et al. (2025), Digitalisation, artificial intelligence and algorithmic management in the workplace: Shaping the future of work, European Parliamentary Research Service, Brussels.
[4] Gonzalez Vazquez, I. et al. (2025), “Digital Monitoring, Algorithmic Management and the Platformisation of Work in Europe”, No. JRC143072, Publications Office of the European Union, Luxembourg.
[3] Milanez, A., A. Lemmens and C. Ruggiu (2025), Algorithmic management in the workplace: New evidence from an OECD employer survey, OECD Artificial Intelligence Papers, No. 31, OECD Publishing.
[2] OECD (2025), Recommendation of the Council on Artificial Intelligence, OECD/LEGAL/0449, https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449.
Contact
Anna MILANEZ (✉ anna.milanez@oecd.org).
This policy brief contributes to the OECD’s Artificial Intelligence in Work, Innovation, Productivity and Skills (AI-WIPS) programme, which provides policymakers with new evidence and analysis to keep abreast of the fast-evolving changes in AI capabilities and diffusion and their implications for the world of work. The programme aims to help ensure that adoption of AI in the world of work is effective, beneficial to all, people‑centred and accepted by the population at large. AI-WIPS is supported by the German Federal Ministry of Labour and Social Affairs (BMAS) and will complement the work of the German AI Observatory in the Ministry’s Policy Lab Digital, Work & Society. For more information, visit https://oecd.ai/work-innovation-productivity-skills and https://denkfabrik-bmas.de/.
← 1. The high overall adoption rates reflected in the OECD figures is due to the breadth of the definition of algorithmic management used. Prevalence figures for specific types of algorithmic management tools were also gathered and can be found in (Milanez, Lemmens and Ruggiu, 2025[3]). Other valuable studies present evidence on adoption rates gathered via employee surveys, such as (Gonzalez Vazquez et al., 2025[4]).