When Machines Sack Humans: The Ethics Of AI Terminations

In warehouses, delivery services, and low-wage retail jobs across the UK, an unsettling trend is taking root: workers are being dismissed not by managers, but by machines. From missing performance targets to unexplained data anomalies, the reasons behind these terminations are often buried within algorithmic decision systems that operate without human oversight. As artificial intelligence becomes more embedded in workforce management, it is raising profound ethical concerns—chief among them, what it means to be fired by a machine.

The deployment of AI in human resources has accelerated rapidly, especially in industries driven by speed, scale, and standardisation. Employers promise efficiency, consistency, and reduced bias. But the lived experience of workers suggests a different reality—one where opacity, unfairness, and dehumanisation are increasingly common. At the heart of this shift lies a difficult question: should an algorithm be allowed to end someone’s employment?

AI Is Managing—and Terminating—Real Jobs

AI systems are already handling a growing share of HR processes: scheduling shifts, evaluating performance metrics, monitoring location data, and enforcing compliance with internal rules. In some cases, these systems now go a step further—triggering warnings, suspensions, or even automatic termination notices based on algorithmic assessments.

This is particularly prevalent in sectors like logistics, ride-hailing, food delivery, and retail—industries where workers are often treated as independent contractors or casual staff. These platforms rely heavily on performance tracking tools that feed into automated decision systems. Workers may receive no prior warning or explanation before being “deactivated” from the system, often learning only after the fact that they’ve been flagged for supposed underperformance or policy breaches.

Ethical Pitfalls of Algorithmic Terminations

The use of AI in employment decisions raises a host of ethical issues—many of which stem from the very nature of how these systems operate.

1. Lack of Transparency

Many workers are never told how their data is used or what thresholds trigger disciplinary action. Algorithms process vast datasets—ranging from GPS movements to task completion times—but provide little insight into their reasoning. The result is a deeply opaque system where workers cannot contest or even understand the grounds for dismissal.

2. No Room for Context or Compassion

Unlike human managers, AI lacks the ability to contextualise performance. A missed shift due to illness, technical errors in an app, or location mismatches caused by software bugs may all be interpreted as “non-compliance.” In the absence of human review, such decisions can be both inaccurate and unjust.

3. Built-in Bias

AI systems are trained on historical data. If that data reflects discriminatory patterns—whether based on race, gender, or work history—those patterns can be reinforced. In low-wage sectors, where monitoring is often more intense for certain groups, the risk of algorithmic bias is particularly acute.

4. Undermining Human Dignity

To be dismissed by an automated process, without explanation or the opportunity to appeal, strips away a fundamental aspect of employment: recognition of personhood. Decisions that affect one’s income, identity, and wellbeing should not be reduced to an impersonal calculation.

Disproportionate Impact on Vulnerable Workers

The ethical issues are compounded by the fact that these systems tend to target the least empowered. Low-paid, casual, and gig workers—already operating with minimal job security—often lack the knowledge, resources, or legal standing to challenge AI-led decisions.

Power asymmetries are central to the problem. Employers can point to automated systems as neutral, sidestepping accountability. Workers, meanwhile, are left to navigate an invisible decision-making process with no clear appeals mechanism.

UK law currently offers limited protection in these scenarios. Many workers affected by algorithmic decisions are not formally employed, meaning they do not benefit from unfair dismissal protections. Even for those who are, the law is still catching up to the realities of algorithmic management.

How Other Jurisdictions Are Responding

While the UK remains largely reactive, other countries are beginning to confront these issues more directly.

The European Union’s proposed AI Act designates employment-related AI as “high-risk,” mandating transparency, human oversight, and accountability. It would require companies to explain how decisions are made and allow individuals to contest automated outcomes.

Some EU member states have gone further. Spain and the Netherlands, for instance, now require platforms to inform gig workers of how algorithmic decisions affect their employment. These moves signal a growing recognition that employment law must adapt to the algorithmic age.

What Ethics Demands

There is growing consensus around a set of principles for ethical AI use in the workplace:

• Transparency: Workers must be informed when AI is used to make employment decisions and understand how those decisions are made.

• Accountability: There must be clear lines of responsibility. Employers cannot hide behind “the system.”

• Human Oversight: Automated decisions should be reviewed by a human before they take effect, particularly in high-impact scenarios like termination.

• Fairness: Systems must be audited for bias, and workers must have the ability to challenge outcomes.

Crucially, ethical frameworks stress the need for human-in-the-loop systems—where machines support decision-making but do not replace it. This is particularly important in cases of dismissal, where stakes are high and decisions cannot be reduced to simple metrics.

A Call for Change

The rise of AI-driven employment decisions is not inherently unethical—but it becomes so when implemented without safeguards, transparency, or accountability. At a minimum, UK regulators need to update employment protections to reflect new technological realities. This includes expanding worker rights to explanation and appeal, clarifying employer responsibilities, and introducing clearer standards for AI usage in HR.

Trade unions and civil society groups also have a role to play. From pushing for transparency laws to supporting legal challenges against wrongful dismissals, organised labour must adapt to this new front in workplace rights.

Conclusion

AI has the potential to make HR decisions faster and more consistent. But when it’s used to sack workers—especially those with the least protection—it raises fundamental ethical concerns. Dismissing a person is never just a data problem; it’s a human one. Until regulators, employers, and technology providers recognise this, the risk of injustice will only grow.

Author: Brett Hurll