Introduction

In this Gen AI era, with plenty of content exploding online, we may have the feeling of “deja vu” that we spend so much time reading content; it seems like making sense at first glance, but you start to realize the question is left unexplained when you sit with it, and you get nothing memorable after you read it.

As of 2025, a report by the Graphite team shows that 50% of the online content is AI-generated. The unrestrained usage of AI is leading to Workslop, destroying the team’s productivity and hurting the business’s credibility. The inaccuracy of AI-generated content may potentially lead to poor decisions, reduce customer experience, and increase operational costs. Before we explain everything in detail, let’s learn about why AI-generated content cannot be completely trusted.

Model Collapse

Model collapse, a concept detailed in a Nature article by a team of researchers led by Ilia Shumailov, is a process in which AI-generated content contaminates the training data set of the next generation, and this recursive process causes the inaccuracy of representing reality.

We should pay attention to this phenomenon because AI-generated data may not be 100% match to real-world conditions. If you rely too much on AI to predict market trends, it will lead to poor decisions. If your content or campaign doesn’t match your customers’ expectations, it will reduce your CX, wasting your money and hurting your credibility in the long run.

Why should we care?

The Model Collapse is a fact existing in AI data training. If the concerns above are potential risks, here are the problems that have already happened in business.

Workslop

AI slop means the low-quality AI-generated posts, as defined in this HBR article. When the AI slop is applied to your business workflow, slowing the business operating efficiency, it is referred to as “workslop,” meaning the work that looks fine at a glance but lacks the quality to advance a given task.

The crucial issue is that the responsibility is shifting from the creator to the receiver. It may take creators only a few minutes to generate a report, but receivers may spend another hour qualifying the content, which originally should be handled by the creator.

According to the data revealed in this HBR article, employees who encounter the workslop estimate that 15.4% of the work is qualified. Also, after receiving the AI-generated work, they will perceive their colleagues as 54% less creative, 50% less capable, 49% less reliable, 42% less trustworthy, and 37% less intelligent.

We should be aware that the usage of Gen AI without clear rules and boundaries has already decreased the team productivity, from workflow to colleague trust. If this is what Gen AI brought to your team management, then the following problem is what it brings to your business results.

Hurt your SEO and credibility

As the Google SEO documentation mentions, their system rewards quality content. The simple version of how SEO works is that when your website receives more traffic, it can indicate to Google that your site is popular and relevant to users, leading to a better user experience, which is a factor Google considers in its ranking algorithm. 

The high volume of AI slop will lower your content quality and trigger Google’s spam filter. Meanwhile, you waste your users' time gathering meaningless, inaccurate information, which leads to lower user experience and engagement. Then they will start to doubt your reputation, resulting in low web traffic. Therefore, in the long term, it hurts your credibility and SEO ranking.

If the SEO ranking is a technical result that seems far away from your business's credibility, then the article published in FIU reviews the direct perception. Creators who disclose the use of AI will be perceived negatively by people, regardless of their existing reputation.

The raw examples and data revealed by other professionals may not be authentic enough. Here is a typical example of my experience with Gen AI.

Real-life example

During my 2nd draft of this article, I used ChatGPT to find a few credible sources to support my thesis. It introduces me to a research paper and affirms that it will support the previous “Hurt your SEO and credibility” section. However, it took me another hour to read the research paper and discover that it proves AI-generated content is polluting the entire web search ecosystem, leaving heavy maintenance tasks for retrieval strategists; yet, it doesn’t align with my objective.

This kind of confirmation job has filled up my workflow for 3 years, and this is what Gen AI brings you. You spend 30 seconds writing a prompt, and AI uses another 1 minute to generate a bunch of unchecked data in an affirmative voice that confuses you. Most people will believe the generated content is true, but in reality, no matter how affirmatively the AI assures you, its authenticity remains unchecked. 

AI creates illusory confidence for a result that sounds right, leading to wrong citations. Also, it increases your cognitive load in checking its data. 

Thus, how do we handle the tool properly to avoid all these problems brought by Gen AI? Here, I offer a few suggestions to adopt Gen AI properly.

How to integrate AI properly

Applying Gen AI practically in business workflows is a trend, but it also faces some challenges. It is worth another article to introduce how to smoothly achieve AI adoption from a more critical and rigorous perspective. But with our experience trying to integrate Gen AI into our workflow, which has not brought us any significant results, here are some practical suggestions I can offer:

Mindset

As the leader or a significant member of your team, you should position yourself as a creator and decider, and integrate AI as an assistant. This means that AI is just a tool to provide data for you to consider. Your work is the foundation, and you decide whether to let AI iterate on the work. 

Clear guidelines

If you’re a leader of your team, set clear guidelines and boundaries for how your team is going to use the AI. Before submitting any AI-generated work, everyone should double-check if it is qualified to merge into the main branch of your team's workflow, avoiding delegating the workslop to another teammate. A primary principle is forbidding just giving a prompt to AI and letting it create the content.

An article posted on UF News reveals that employees can often detect the AI-generated content. Only 40% - 52% of employees will see supervisors as sincere when they use AI, compared to 83% for low usage. Additionally, 95% of employees perceive low-AI supervisors' messages as professional, compared to 69-73% for highly reliant.

This indicates that in a relationship-oriented conversation or message, the leader should carefully consider whether to use AI. Keeping the team cohesive is a win-win result for both personal and business.

Small chunk experiment

Gen AI is good at a small chunk of work. If you let them fix your grammar, write a short chunk of code, find a precise word, or teach you how to search information more effectively. It will definitely accelerate your process in coming up with ideas. Also, these explorations should stay in personal workflow and be integrated into the team document once they pass the personal experiment.

Conclusion

In conclusion, the unpredictable inaccuracy of AI-generated content will lead to poor decisions. If your team uses AI recklessly, it will reduce trust among colleagues, increase operational costs, and hurt your business's credibility.

AT, as a web agency, we follow the principles from which you can get inspiration:

  1. Don’t use or at least don’t rely on Gen AI for any production-level work to ensure accountable results.

  2. Take your time to create original and high-quality content for business sustainability.

  3. Validate and iterate on your work that encounters AI-generated content to avoid ineffective team communication.

  4. Don’t use AI for relationship-oriented conversations to ensure the team's trust.

  5. Only use AI on a small chunk of work that does not influence the operational efficiency.

Following these principles, which AT will also help you with, will ensure a more reliable business environment and better long-term performance.