Automation and algorithmic management are reshaping every corner of the labour market. Drawing on Sarah O’Connor’s new book We Are Not Machines, this post unpacks how workers are bending to fit the needs of artificial intelligence, why that matters and what can be done to build a healthier relationship between people and technology.
The Central Argument
O’Connor contends that the real revolution is not robots replacing humans outright but humans adapting themselves to serve the logic of software. Whether it is warehouse pickers “optimised” by handheld scanners or call-centre staff tied to sentiment-analysis dashboards, the power dynamic has flipped: people now accommodate machines, not the other way around.
Three Core Themes
1. Algorithmic Management: Schedules, targets and evaluations are increasingly generated by opaque systems. Workers rarely see the data that judges them, which limits their ability to contest unfair ratings or errors.
2. Fragmented Work Identities: Gig-economy platforms treat individuals as interchangeable task units. The resulting churn erodes career progression, weakens bargaining power and blurs the boundary between “on” and “off” hours.
3. The Human Cost: Intensified surveillance and unattainable performance metrics translate into stress, musculoskeletal injuries and a sense of disposability. The psychological load is often masked by the promise of flexibility and entrepreneurial freedom.
Why This Matters
If current trends continue, entire sectors could normalise machine-centric workflows that undervalue creativity, discretion and empathy. That risks:
- Stagnant wages as piece-rates suppress income growth
- Widening inequality between those who design systems and those subject to them
- Democratic deficits when corporate code effectively writes workplace law
Pathways Forward
Regulation and Transparency
Mandate algorithmic impact assessments, require explainability for automated decision-making and grant workers access to their own data. The EU’s upcoming AI Act hints at this direction but enforcement will be critical.
Worker Voice and Collective Bargaining
Trade unions and new-model collectives can negotiate algorithmic guardrails—caps on real-time monitoring, limits on shift volatility and coexistence plans when new tech is introduced.
Designing Humane Systems
Product teams can apply human-in-command principles: augment, don’t automate; maintain a clear channel to override the system; and incorporate wellbeing metrics alongside productivity KPIs.
Takeaways for Different Audiences
Policymakers: Treat data rights as labour rights. Consider portable ratings that workers can carry across platforms.
Employers: Embed cross-functional ethics reviews when deploying AI tools. Measure success not only by output but by retention and satisfaction.
Workers: Document algorithmic inconsistencies, share information and push for transparency through formal and informal networks.
We Are Not Machines is ultimately a call to re-humanise the future of work. Technology should relieve drudgery, not redefine people as flexible extensions of code. By insisting on transparency, participatory design and robust labour protections, we can shape AI systems that serve human flourishing rather than stunt it.
