Can an algorithm fairly decide who gets laid off?

A group of 26 former Meta employees is suing the company, claiming they were unfairly targeted for layoffs. The employees allege that Meta used a network of internal software—which they refer to as AI tools—to rank performance and create lists of people to fire. The core of their complaint is that these tools did not ignore periods where employees were on protected parental or medical leave. Consequently, when the AI ranked people based on productivity or internal activity, it essentially penalized those who hadn't been working during their leave. By failing to account for this legally protected time off, the plaintiffs argue, the software disproportionately selected them for termination. Meta has responded by stating that their layoff decisions were made by humans, not by AI.

To understand how this happens, think of the AI as a high-speed accountant. These tools are fed vast amounts of data—things like how many projects someone finished, their internal communication patterns, or how much they engaged with various company dashboards. An algorithm is just a set of instructions that processes this data to turn it into a score. The problem arises when the person building the program forgets to add a rule that says these specific time gaps shouldn't be counted against the person. If you feed the system months of records showing zero activity because someone was home with a newborn, the algorithm doesn't know that's a protected, legal choice. It only sees a low output number. In its logic, a low number is simply a low number, leading it to suggest that person for a layoff. The human managers relying on these scores may never realize they are looking at flawed math until the damage is done.

This case highlights a major trap in modern corporate management. When a company uses software to make decisions, it’s easy to treat the result as an objective truth. However, software is only as fair as the rules and data we give it. If leaders trust these tools too much, they risk outsourcing their moral and legal responsibilities to an algorithm that doesn't understand context. As we continue to integrate these tools into professional life, the question isn't just whether the AI is accurate, but who is being held responsible when the math doesn't account for the humans behind the data.