A backlash against artificial intelligence is gaining momentum. What once seemed like a clever shortcut to automate routine tasks now often feels like a drag on performance. A new report defines this problem as 'workslop' – AI-generated content that appears polished but lacks accuracy, substance, or proper review. Nearly half (45%) of US professionals say workslop has made them more cautious about using AI in the workplace, according to a survey of professionals.

The report highlights the top risks associated with workslop: lower trust in AI (57% of respondents), reduced productivity (51%), and damage to a company's reputation (46%). For a technology designed to boost efficiency, these findings present an uncomfortable reality: AI is reshaping how work gets done, but not always for the better. The challenge lies in moving from blind adoption to smart, human-centered deployment.

Understanding workslop and its consequences

Workslop arises when employees rely on generative or agentic AI without sufficient oversight. Tools produce text, code, or summaries that sound convincing but contain errors, hallucinations, or irrelevant content. When these outputs are used without human review, they introduce errors into workflows, forcing people to spend extra time correcting mistakes – negating any efficiency gains.

For example, a team might use an AI meeting summarizer that misses key decisions or misinterprets action items. Instead of saving time, the team must cross-check summaries against recordings, causing delays. Over time, such experiences erode confidence in AI and lead to broader skepticism. The report found that 57% of professionals now trust AI less after encountering workslop, and 46% worry about reputational harm if low-quality outputs reach clients or partners.

Step one: Rethinking productivity

Business leaders who have successfully integrated AI advocate for a fundamental shift in how productivity is defined. One chief technology officer of a global content and technology company explains that an 'AI-first, human-second' work pattern is emerging. Instead of starting a task from scratch, professionals should let AI generate a first draft or initial analysis, then apply their judgment, intuition, and domain expertise to refine the output.

This approach changes the nature of work from execution to curation. In software engineering, for instance, developers now use AI to generate boilerplate code, then focus on testing, optimization, and architecture. Similar transitions are expected in other fields. The key is distinguishing between tasks where AI adds genuine value – such as data summarization, pattern recognition, or routine documentation – and tasks where human creativity, empathy, or strategic thinking is irreplaceable.

A CIO at a technology specialist company has developed a model to evaluate which AI tools deliver real productivity gains. The model examines multiple vectors: business risk, financial return, and time saved on meaningful activities. For example, generating notes from a meeting that nobody will read does not create value. But an agent that automates supplier invoice reconciliation – cutting a three‑hour process to 20 minutes – is a clear win. Organizations must create a culture where employees are trained to recognize workslop and choose tools wisely.

Another CIO, from a property company, emphasizes the importance of a learning culture. He recommends that organizations openly discuss the risks of workslop and celebrate cases where AI augments, rather than replaces, human expertise. 'We need to understand what AI cannot do – it cannot inspire people or create genuinely novel ideas because it is inherently recursive. That is where human judgment remains essential,' he says. By framing AI as an assistant to educated professionals, companies can avoid the productivity trap.

Step two: Being persistent

Implementing AI is only the beginning. Real returns require tenacity. The same chief technology officer notes that many early adopters gave up when AI tools did not work perfectly out of the box. They turned off the technology and declared it unready, missing the chance to refine and adapt. In contrast, the most successful teams built custom workflows around the AI: grounding it with curated data, writing clear prompts, and iterating based on feedback.

In one instance, a single curious individual within a team invested extra time to hone the AI's performance. That person's efforts eventually lifted the entire team to a new level of efficiency. The lesson: persistence transforms AI from a novelty into a reliable productivity multiplier. Leaders should encourage experimentation and reward those who invest effort in fine-tuning tools.

The CIO of the technology specialist says that employees who master the blend of AI and human expertise will become highly sought after. They will also expect their future employers to provide advanced AI capabilities. This creates a talent attraction loop: companies that offer smart, well-integrated AI tools will retain and attract top performers. The same CIO warns that professionals are beginning to judge employers based on whether they have agents that automate tedious tasks – or leave workers stuck with workslop.

A property company CIO agrees that persistence will pay off. Despite rising backlash, AI is not going away. The technology will continue to evolve, and professionals who learn to exploit its strengths while guarding against its weaknesses will remain competitive. The message is clear: do not abandon AI at the first sign of workslop. Invest time in training, prompt engineering, and output review. Build feedback loops so that the system learns from human corrections.

Practical steps for individuals and teams

For professionals looking to avoid workslop, several actionable tactics emerge from the expert insights. Start by auditing which tasks actually benefit from AI. Use the 'AI-first, human-second' pattern: let the tool produce a draft, then allocate time for verification and improvement. Adopt a persistent mindset – if the first output is poor, refine the prompt, add context, or try a different model. Keep a log of successful prompts and share them with the team.

Organizations should establish an internal AI marketplace where employees can choose vetted tools with clear performance metrics. Create a 'safe sandbox' where teams can experiment without fear of damaging operations. Provide training on critical evaluation of AI outputs, including how to spot hallucinations and biases. Finally, measure productivity not just by output volume but by the quality and time savings after human oversight.

The debate about when the AI bubble might burst continues, but those on the ground see that the technology is here for the long haul. Workslop is a symptom of immature deployment, not a reason to reject AI entirely. By rethinking productivity and persisting through the learning curve, professionals can turn the tool into a genuine advantage – one that boosts, rather than drains, their daily effectiveness.

Source: ZDNET News