The economic cost of AI workslop
Nearly all of us have experienced a self-reinforcing problem. We get a bad night’s sleep and, depleted of energy, rely on coffee to power through our to-do lists. But that final cup — meant to get us to 5 p.m. — comes back to haunt us, and we end up staring at the ceiling for hours at bedtime.
The process repeats until we finally cut our afternoon caffeine jolt for good. As examples like this demonstrate, there is a reason for the term ‘vicious cycle.’ Cause-and-effect loops are challenging to break.
AI workslop is another example. The Scientific American defines it as “mass-produced, low-quality content.” Fast Company refers to it as the “collective shorthand for digital garbage.” While definitions vary, experts agree: AI workslop is widespread. Unfortunately, the fix for stopping it is not as simple as reducing coffee intake.
Navigating the influx of slop on the internet will require significantly more effort, with workplaces feeling the potential brunt.
Harper’s Magazine reports that more than half of the internet is now generated by bots, creating an “AI feedback loop, corrupting the very real and ‘very true’ human data that was supposed to, in aggregate, make the technology so powerful.”
In short, when AI models train on garbage — in this case, slop — the quality of the output we receive continually degrades. Harvard Business Review reports that many companies are already feeling the effects. Rather than enhancing efficiencies, this low-quality content is hampering productivity. Research shows it creates additional “cleanup work” for employees — an average of two hours per instance — and can foster resentment among coworkers. It is a pervasive problem, with MIT Media Lab finding “95% of organizations [now] see no measurable return on their investment” from AI.
Take, for example, if a manager asked an employee to write a report about a business line to gauge its success. The individual, short on time, decides to upload data to an AI tool to draft the initial content. The output is light on substance, but it reads so well that the employee sends it without any edits. The manager, with a dozen reports to review, then asks an AI platform to summarize the findings before uploading the still-bloated document to a shared drive. Did leveraging AI tools help the company identify potential pain points for operational improvements? No. In fact, it likely wasted time and internal resources for little value.
What’s the lesson to learn? Instead of blanket adoption, businesses should be smart and strategic about when and how they enable employees to use AI tools. As Forbes points out, the “antidote [to slop] isn’t more AI,” but, instead, to use the technology to “extend what’s already strong” in our workplaces. Companies should conduct an internal assessment of their needs and goals to determine where AI tools can provide real value to their processes.
For example, could AI help perform repetitive or routine tasks, so employees can focus on more complex or nuanced responsibilities where they have expertise? If companies decide to deploy AI tools, they should establish and implement comprehensive governance policies, ideally with the help of an experienced cybersecurity professional, to ensure that all employees receive adequate training on how to use vetted tools effectively and safely.
Recent headlines have declared that AI slop is “destroying productivity” and “muddying the American workplace.” But AI isn’t all bad. The technology can help enhance efficiency and extend our capacity, including in the workplace. The onus is on us to utilize it responsibly if we want to help break the vicious cycle of workslop.
Editor’s note: Chris Wright is co-founder and partner at Sullivan Wright Technologies, an Arkansas-based firm that provides cybersecurity, information technology and security compliance services. The opinions expressed are those of the author.