Research worth reading
What research says about people working with AI and robots, and about what makes teams work. Each link goes to the original publisher.
People, AI and robots
Nature Human Behaviour (Vaccaro, Almaatouq & Malone, MIT) — open access · 2024 · PDF
When combinations of humans and AI are useful: a systematic review and meta-analysis (opens the publisher site in a new tab) ↗
Across 106 experiments, human–AI combinations on average did worse than the best of humans or AI alone — with losses in decision tasks and gains in content-creation tasks.
NBER Working Paper 31161 (Brynjolfsson, Li & Raymond) · 2023 · PDF
Generative AI at Work (opens the publisher site in a new tab) ↗
An AI assistant raised customer-support productivity by 14% on average and 34% for novice and low-skilled agents, with minimal impact on the most experienced.
OECD · 2023 · PDF
The impact of AI on the workplace: Main findings from the OECD AI surveys of employers and workers (opens the publisher site in a new tab) ↗
Workers and employers in finance and manufacturing are generally very positive about AI’s impact on performance and working conditions; training and worker consultation go with better outcomes.
Stanford HAI · 2025 · PDF
Artificial Intelligence Index Report 2025 (opens the publisher site in a new tab) ↗
A growing body of research confirms that AI boosts productivity and, in most cases, helps narrow skill gaps across the workforce.
NIST · 2023 · PDF
Artificial Intelligence Risk Management Framework (AI RMF 1.0) (opens the publisher site in a new tab) ↗
Human roles in decision-making and oversight of AI need to be clearly defined; human–AI configurations range from fully autonomous to fully manual.
ILO Working Paper 96 (Gmyrek, Berg & Bescond) · 2023 · PDF
Generative AI and Jobs: A global analysis of potential effects on job quantity and quality (opens the publisher site in a new tab) ↗
Only clerical work is highly exposed; for most occupations generative AI is more likely to augment work — automating some tasks while leaving time for other duties — than to automate whole jobs.
MIT Task Force on the Work of the Future · 2020 · PDF
The Work of the Future: Building Better Jobs in an Age of Intelligent Machines (opens the publisher site in a new tab) ↗
Robotics and AI are reshaping industries, but their effects on work unfold gradually — time to build skills and steer change toward broad benefit.
EU-OSHA (European Agency for Safety and Health at Work) · 2022 · PDF
Advanced robotics and automation: implications for occupational safety and health (opens the publisher site in a new tab) ↗
Robots, including cobots, can take over non-ergonomic and monotonous tasks, freeing people for more cognitive or creative work — with physical, psychosocial and organisational effects to manage.
Collaboration: the fundamentals
Psychological Science in the Public Interest (Kozlowski & Ilgen) — free access · 2006
Enhancing the Effectiveness of Work Groups and Teams (opens the publisher site in a new tab) ↗
A review of 50+ years of team research: cognitive, motivational and behavioural team processes drive effectiveness, and levers such as composition, training and leadership can improve them.
Google re:Work (People Analytics) · 2015
Understand team effectiveness (Project Aristotle) (opens the publisher site in a new tab) ↗
Studying 180 Google teams, researchers found that how a team works together matters more than who is on it — led by psychological safety, then dependability, structure and clarity, meaning, and impact.
PNAS (Riedl, Kim, Gupta, Malone & Woolley) · 2021
Quantifying collective intelligence in human groups (opens the publisher site in a new tab) ↗
Across 22 studies with 5,279 people in 1,356 groups, a group’s collective intelligence was most strongly predicted by its collaboration process — more than individual skill or group composition.
U.S. National Research Council (National Academies) · 2015 · PDF
Enhancing the Effectiveness of Team Science — Report Brief (opens the publisher site in a new tab) ↗
Team science can deliver greater impact, innovation, productivity and reach than solo work — but diversity, deep integration, size and misaligned goals create challenges teams must manage.