Working conditions at tech AI startup Mercor are under-scrutiny, with ex-employees speaking to media under condition of anonymity describing "hostile" and "inhumane" working conditions for the 30,000 data labellers powering the company. Untenably long working hours, excessive targets and pressure to deliver at speed have reportedly led to employees resigning within one week, without bonuses or equity. One employee who did so cited "inhumane working conditions".

While the company stated that practices such as tying bonuses to performance are industry standard, the extent to which issues aroud deception at hiring, precarious roles and consistent team restructuring have adversely impacted employees remained unaddressed.

An analysis in Biz Tech Weekly highlights how the gig economy has shifted to employing workers with PhDs, but highlights how one fifth of Mercor employees are homeless, "many" depend on public financial assistance, most cannot cover basic living costs, and a high turnover and abrupt dismissals are a feature of the company's operating model. It states that responsible AI cannot be separated from broader approaches to responsible sourcing, with data labellers in the supply chains of big AI businesses facing increasingly unsustainable conditions.