We all know that one person in the office—the boss’s relative—who, despite their frequent missteps and unpredictable output, is a permanent fixture. Start thinking this way about generative AI.

Much like the boss’s relative, it can often be helpful but can just as easily be oblivious and submit work riddled with inaccuracies. Yet, here we are, stuck with them, figuring out how to make it work.

Let’s talk about Alex, our generative AI equivalent of the boss’s well-connected relative. Alex has a stellar educational background, having attended the best schools. Theoretically, Alex should be a rockstar. But in practice? Well, let’s just say Alex often misses the mark.

The allure of generative AI is undeniable. Who wouldn’t want a tool that can seemingly handle customer service, content creation, and even complex problem-solving? But the reality is that these AIs can be as unpredictable as Alex on a Monday morning after a long weekend. It can be like playing a game of roulette every time you hit ‘generate.’

Just like dealing with Alex, you need to develop strategies to manage and mitigate the risks. Here’s how you can deal with it:

1. Double-Check and Verify: Always verify the information provided by generative AI. Treat it as a starting point, not the final word. Because let’s be honest, would you trust Alex’s first draft without a second look?

2. Provide Clear and Specific Prompts: The quality of an AI’s output depends on how well you guide it. Think of it as giving Alex a very specific task—less room for error that way.

3. Implement Safeguards: Use automated tools and human oversight to catch and correct errors. Have workflows that incorporate reviews and approvals, especially for critical content. Because, let’s face it, Alex isn’t known for their attention to detail.

4. Continuous Learning: Stay informed about the strengths and limitations of generative AI. Keep up with updates and improvements to get the most out of it. After all, even Alex has the potential to improve with the right guidance.

5. Set Realistic Expectations: Understand that generative AI is a tool, not an oracle. Setting realistic expectations about its performance can help mitigate frustration and ensure a more productive relationship. Remember, it’s like managing Alex—expecting miracles will only lead to disappointment.

The boss’s relative is going to keep their job no matter what your opinion on the matter. Ditto generative AI, so you might as well figure out how to work with them.

While you’re here, poke around this blog and site, and you’ll see why I talk about these things, what I think, what I do to help. You might also be curious about LawQi.