WHAT EMPLOYERS SHOULD DO TO AVOID AI SLOP The good news is that a few simple steps will help you use AI without creating slop. Here are five steps you can take to set guardrails that preserve human judgment, corporate accountability and quality of work. 1. Define “Acceptable AI Use,” Not Just “Permitted AI Use” The first generation of corporate AI policies, released in the wake of ChatGPT’s release and the subsequent explosion of GenAI (roughly 2023-2024), is already outdated. If you look at your company policy, it probably focuses on “AI is allowed for X.” But that’s not enough these days. Instead, you need to go further and spell out: • Where AI can assist and where it needs to be avoided • Where human judgment must lead • That AI output must always be reviewed before use 2. Assign Ownership for AI-Assisted Work Every AI-generated document should have a human owner responsible for ensuring accuracy, tone and consistency with the company’s values and policies. 3. Train Managers to Spot Slop Managers and content approvers don’t need to be AI experts, but they do need to recognize some of the telltale signs of AI usage so it can be edited out before release. • Overly generic language that could apply to any role, department or company • Perfectly structured paragraphs that lack a human voice, specificity, or the natural friction that comes along with human writing • Excessive use of em-dashes like this — often multiple per paragraph — to create the illusion of nuance • “It’s not just X — it’s Y” construction, used repeatedly to inflate ordinary points into faux insights (like “It’s not just a change — it’s a revolution”) • False confidence in incorrect or oversimplified statements, especially about legal, technical or operational issues • Repetition of the same idea using slightly different phrasing rather than adding new substance • Buzzword stacking (“synergy,” “alignment,” “value creation,” “impactful outcomes”) without concrete examples • Vague conclusions that gesture toward action without assigning ownership or next steps 4. Slow Down High-Risk Uses For legal, HR, compliance and external communications, make sure you require human review before anything is delivered. It will be almost impossible for AI to generate sound communications in these areas without human customization, often a great deal of it. 5. Measure Outcomes, Not Output You will want your organization to shift performance conversations away from questions like “How fast was this produced?” and “How many pieces of content did you develop?” Instead, consider whether the communication effectively solved the problem and reduced the need for follow-up questions. FP IS ALL IN ON AI — BUT THE RIGHT KIND OF AI Don’t get us wrong. Fisher Phillips is a massive supporter of AI. We host an annual AI Conference that brings together thought leaders from across the country. We hold dozens of webinars each year about AI use (including our popular AI Forums). We’ve published hundreds of AI-related Insights, have an active AI, Data, and Analytics Team, and offer products and services to help businesses master the use of AI. We understand the value it can bring to businesses and corporate leaders. But (with apologies to Spider-Man’s Uncle Ben), with great power comes great responsibility. By following the steps we’ve outlined, you can responsibly use AI without contributing to the scourge of slop. 12 SAN DIEGO DEALER
RkJQdWJsaXNoZXIy MTg3NDExNQ==