AI in the Workplace: Where It Helps and Where It Still Falls Short (Chris Wright Commentary)

(Shutterstock)

Is artificial intelligence (AI) the saving grace for any workplace challenge? Do you need a professional infographic for a slide deck? Whip it up in ChatGPT. Are you feeling stumped when writing a blog post or client newsletter? Do a quick query for inspiration. But what about more complex, high-stakes tasks like delving into cancer mortality rates to inform prevention efforts?

If the database is pristine, with no gaps or flaws, it’s possible that AI could help with the initial run-through. However, the technology could quickly run into issues if confronted with nuanced cases, such as patients with comorbidities like obesity or high blood pressure. To be effective, AI requires explicit data points. Without these figures, it could give an inaccurate view of existing cancer mortality patterns—potentially thwarting the effectiveness of mitigation strategies. 

Today, AI is often portrayed as a default solution, capable of helping businesses and their employees accomplish nearly any on-the-job feat. But just because AI can do certain things doesn’t mean it will do them well. We must recognize the difference, considering what AI is good for—and where it’s less beneficial—so we can harness its power for our collective benefit.

Where AI excels: AI can quickly comb through big data sets to identify patterns. Take, for example, static code scanning. AI can simultaneously look through millions of lines, making contextual connections across the entire code base to pinpoint potential vulnerabilities. The same task would require a human an exorbitant amount of time and energy. AI is also adept at performing repetitive, routine tasks, which can significantly improve companies’ internal capacity and workers’ productivity. By shifting these functions to AI, employees have more time to focus on complicated or urgent issues that require human brainpower.

Where AI falls short: Nearly everyone has heard the expression, “It must be true. I heard it on the internet.” Unfortunately, the internet is filled with incorrect information, some malicious. AI hallucinations occur because large language models, such as ChatGPT, generate content from what’s available online. If data exists publicly, LLMs can access and use it to create results. Generative AI doesn’t know and can’t predict what we want. We must be detailed and diligent with our queries to ensure the desired response, and even then, we must still verify the content. 

The bottom line: AI is novel and “cool,” with a growing list of benefits—time savings, workplace efficiencies, creative inspiration… But the technology isn’t the be-all and end-all. Just as we wouldn’t stop teaching our children how to add without a calculator, we shouldn’t rely only on AI. Overdependence on any technology—AI or otherwise—can lead to adverse long-term effects, chiefly a lack of innovation, creativity and strategic thinking. As AI expands in capabilities, we must recognize its shortfalls, alter our approaches and, sometimes, adjust our expectations to ensure maximum workplace benefits.

Christopher Wright is co-founder and partner at Sullivan Wright Technologies, an Arkansas-based firm providing tailored cybersecurity, IT and security compliance services.