Can AI Fire Me? What You Need to Know About Automated Employment Decisions

How AI gets involved in termination decisions

AI rarely fires anyone directly — but it increasingly influences the decisions that lead to termination. The most common pathways are: algorithmic performance management that generates performance ratings used in performance improvement processes; AI-driven scheduling or task allocation that creates documented underperformance where the AI-set targets were unrealistic; automated misconduct detection in monitoring or surveillance systems that flags incidents for disciplinary review; AI-generated redundancy selection scoring that ranks employees for redundancy; and gig platform algorithmic deactivation, where the algorithm simply ceases to offer work.

Each of these raises different legal issues and attracts different remedies. The key question in most jurisdictions is whether there was a fair process with genuine human decision-making and an opportunity to respond.

Unfair dismissal protections: does AI change the process requirements?

AI does not reduce the procedural obligations that apply to dismissal. In Australia, the UK, and the EU, an employer must follow a fair process regardless of what AI recommends or detects.

In Australia (under the Fair Work Act), a dismissal must be for a valid reason and be procedurally fair. Procedural fairness requires: notifying the employee of the reason they may be dismissed; giving the employee an opportunity to respond; and allowing the employee to have a support person present for any discussion about dismissal. An AI-generated performance rating that triggers a PIP and ultimately a dismissal is not itself evidence of a valid reason — the employer must still be able to justify the rating and the decision in a Fair Work Commission hearing. Where algorithmic management generated the "evidence" of poor performance, the reliability and fairness of that evidence can be challenged.

In the UK (under the Employment Rights Act 1996), employees with two or more years' service are protected against unfair dismissal. A dismissal based on AI-generated misconduct detection or performance assessment must still be the product of a reasonable investigation, reasonable belief in the facts, and a process falling within the band of reasonable responses. An Employment Tribunal can examine how the AI system worked, whether its outputs were reliable, and whether the employer exercised independent judgment rather than simply accepting the AI's recommendation.

In the EU, GDPR Article 22 explicitly applies to AI-influenced termination decisions. If an AI system substantially determines the decision to terminate without genuine human review, the employee may have the right to contest it on Article 22 grounds. Many EU member states add additional procedural requirements — Germany requires works council involvement, France requires specific procedural steps for different categories of termination.

Gig platform algorithmic deactivation: the hardest case

Workers on gig platforms — Uber, DoorDash, Amazon Flex, Upwork — face algorithmic management that can result in deactivation (loss of access to the platform) with little or no explanation. These workers are typically classified as independent contractors, which limits the employment law protections available.

However, the legal landscape is shifting. In Australia, the High Court's 2022 decisions on CFMMEU v Personnel Contracting and ZG Operations v Jamsek clarified the test for employee versus contractor status, and the Fair Work Act's 2023 amendment introduced a new definition of "employee-like" workers with access to the Fair Work Commission. Platform workers may qualify for some Fair Work protections depending on their circumstances. In the UK, the Supreme Court's 2021 decision in Uber BV v Aslam found Uber drivers were "workers" entitled to minimum wage and holiday pay. In the EU, the Platform Work Directive (agreed in 2024) creates a rebuttable presumption of employment for platform workers, shifting the burden to platforms to prove independent contractor status.

If you are deactivated from a platform and believe this was unfair or related to protected activity — such as attempting to organise with other workers — seek legal advice. The law is evolving quickly in this area.

What to do if you think AI led to your dismissal unfairly

Document everything: the AI tool involved, the data it generated, the timeline, communications with your employer, and any concerns you raised about the AI system's accuracy. In Australia, make a subject access request under APP 12 for your personal data including any AI-generated assessments. In the UK and EU, exercise your GDPR Article 15 right of access. Ask your employer to explain the role AI played in the decision. Challenge any AI-generated data you believe is inaccurate. In Australia, file an unfair dismissal application at the Fair Work Commission within 21 days of dismissal. In the UK, apply to the Employment Tribunal within 3 months less one day. Contact ACAS early in the UK for conciliation. In the EU, apply to the relevant employment tribunal or court following your member state's procedure.