When A.I. Creates More Work

A New Workplace Complaint: Cleaning Up the AI

For the past two years, executives have pitched artificial intelligence as a tool that would strip away drudgery, accelerate office work and make companies more efficient. But a growing body of worker testimony is telling a different story: that for many employees, A.I. is not removing tedious labor so much as creating a new kind of it.

The latest complaints are coalescing around what some workers have begun calling “workslop” — A.I.-generated drafts, reports, emails and marketing copy that arrive looking polished and complete, but are riddled with errors, awkward phrasing, false claims or missing context. Instead of saving time, workers say, the material often has to be rewritten from the ground up.

Recent reporting has documented employees describing heavier editing burdens, lower morale and pressure from managers to keep using A.I. tools even when the results create more work. The emerging frustration marks a notable turn in the public conversation over workplace A.I., shifting attention away from corporate promises of productivity and toward the hidden costs absorbed by rank-and-file staff.

That tension is becoming harder to ignore as employers also push A.I. further into another fraught part of working life: hiring.

From Productivity Tool to Cleanup Problem

The complaint about “workslop” is not simply that A.I. makes mistakes. Workers say the problem is that the output often looks convincing enough to move through an organization before anyone notices what is wrong with it. That can leave colleagues downstream — copy editors, analysts, designers, recruiters and managers — to catch inaccuracies and repair the damage.

Survey data frequently cited in the debate suggests the issue is not isolated. Research from BetterUp and Stanford has found that 40 percent of U.S. desk workers said they had received “workslop” in the previous month, with many reporting that they spent meaningful time fixing it. For employers touting efficiency gains, that raises a more uncomfortable question: productivity for whom?

In many offices, the benefits of speed may accrue to the person generating the draft, while the cost is transferred to someone else responsible for ensuring it is usable, accurate and safe to send. Workers say that dynamic can be especially frustrating when management evaluates A.I. by how quickly something is produced rather than by how much labor is required to make it right.

The concern lands at a sensitive moment for companies that have been racing to integrate generative A.I. into everyday workflows. Firms have marketed these tools as assistants for writing, customer service, coding, planning and internal communications. But as adoption spreads, so do reports that the technology is flattening expertise, encouraging volume over quality and making it harder to tell whether a colleague’s work reflects judgment or autocomplete.

Hiring, Too, Is Becoming Automated

The backlash is not limited to current employees. It is also extending to applicants, who increasingly face automated systems before they ever speak to a human being.

A new call for worker accounts of job interviews conducted partly or wholly by A.I. points to a hiring practice that appears widespread enough to generate a steady stream of complaints. The objections go beyond the coldness of talking to a machine. Job seekers have for years described interviews that felt robotic, truncated or impossible to navigate, with fears that automation would fall most heavily on lower-level applicants with the least leverage.

Those concerns have only grown as companies deploy A.I. recruiters and interview systems at greater scale. By the middle of last year, larger news organizations were reporting that employers were using A.I. tools to contact candidates, conduct preliminary interviews and screen applicants in bulk. Supporters say such systems can speed up hiring and reduce administrative strain. Critics argue that they can degrade the candidate experience, obscure how decisions are made and introduce new forms of unfairness under the guise of efficiency.

For applicants, the practical concern is often simple: if an algorithm or automated interviewer misreads a pause, an accent, a disability-related communication style or an unconventional work history, there may be no easy way to contest the result.

Regulators Have Been Warning About This

The labor backlash is arriving against a backdrop of growing regulatory concern. In the United States, officials have already signaled that A.I.-assisted recruiting and screening can create discrimination and compliance risks.

The Equal Employment Opportunity Commission’s enforcement plan for 2024 through 2028 explicitly identifies A.I. and machine-learning tools used in recruiting and hiring as an enforcement priority when they exclude or adversely affect protected groups. New York City, meanwhile, requires employers using automated employment decision tools to provide notice and make public summaries of bias audits.

Those rules do not amount to a single national standard, and the regulatory picture remains uneven. But they underscore that the concerns surrounding automated hiring are no longer theoretical. Employers are being told, in effect, that efficiency cannot come at the expense of transparency, accessibility or civil rights compliance.

What remains less clear is how often companies are meaningfully testing their systems for bias, how much human review exists in practice and whether job seekers understand when A.I. is influencing the process.

Why This Matters Now

The timing is significant. Much recent discussion of A.I. at work has centered on adoption: which companies are buying these tools, how aggressively software makers are selling them and whether businesses can afford to fall behind. The new labor-focused reporting complicates that narrative by asking what happens after the rollout.

If workers are spending hours correcting synthetic output, the promised productivity gains may be overstated or unevenly distributed. If applicants are being screened by opaque systems that feel arbitrary or inaccessible, employers may be trading speed for trust. And if managers insist on A.I. use despite mounting complaints, workplace tension over the technology could deepen.

There are still important unknowns. Anecdotes do not yet show how representative these problems are across industries, or whether some employers are measuring the full cost of cleanup and rework. It is also unclear how often automated interviews materially affect who advances and who is screened out.

But the direction of travel is becoming easier to see. The debate over workplace A.I. is no longer just about what the technology can do. It is increasingly about who bears the burden when it does not work as advertised.

Sources

Further reading and reporting used to add context:

AI News

When A.I. Creates More Work

A New Workplace Complaint: Cleaning Up the AI