Treat artificial intelligence like a stranger, do your own research to verify if what they’re telling you is true and maintain the human perspective when using AI. Those were some of the takeaways from a recent CREW Iowa luncheon that focused on artificial intelligence.

The Oct. 14 event at the Monroe in West Des Moines featured Kira Barclay, founder and managing partner of consulting and data advisory firm Black Antler, as moderator, and Jenny Schmidt, founder of J Schmidt Consulting, and Shalyn Nystrom, principal consultant, data science and analytics, at Source Allies, a technology services company based in Des Moines, as panelists.

Barclay, Schmidt and Nystrom talked about the do’s and don’ts of using AI and gave some advice on how to get started using AI and how to best manage it in the future.

Here are some key takeaways from their conversation. Their comments have been lightly edited for clarity and brevity.

Barclay encouraged users of ChatGPT to ask it why.

“ChatGPT returns an answer. Ask why do you think that? Engage with it like it’s someone sitting next to you. Ask where did you find that information? In our days of doing research on Google, if you found an answer, would you just take that at face value or would you check the source? You check the source. It’s the exact same thing with ChatGPT. What are the sources? Check those sources and understand the context of those sources. If you’re getting advice, you haven’t built that reputation yet, so really fact check and ask more questions like you’re talking to a stranger.”

Nystrom said AI users should always keep asking questions to improve their results.

“The prompt, whatever you’re putting into the chat, it can be massive. There’s so much information that it needs to know to do the job well. If it isn’t exactly what you wanted, ask a few more questions, and once you’re happy, ask it, ‘What can I put in the next time so that we get here in the first place?’ That can help you make your prompts get better and better and you can learn a lot from that. And there’s a lot of training on AI for prompting right now.”

Barclay said the harsh reality of AI is “garbage in, garbage out.”

“What makes these things work is the data that sits underneath it, and it is probably one of the biggest mistakes I see working with organizations today, is underinvesting in the data that sits underneath it.”

Schmidt said it’s important for companies to look at their processes and how they can improve with AI. 

“One of the best influences you can have in your company is looking at your process to figure out how AI can benefit your process but also where you’re not actually following your process, and people are usually what’s causing those data issues, and that’s worth spending some time on because once you get quality data it’s amazing what companies can do. We are all putting data in somehow in the process, so really evaluating what influence you can have can potentially improve that.”

Nystrom said it is critical for people to surround themselves with leaders or partners who are watching for people processing technology to get it implemented.

“It’s huge. Working with various clients, everyone has the same problems, and it’s fun as a consultant to see that I can help a company solve a problem. And none of it has to do with any of the technology that I’m using. It’s all about people and processes and making sure they have the right things in place.”

Schmidt said it’s important to maintain the human perspective.

“ChatGPT is going to be your best friend. It will totally know you. So, just from a human perspective, realizing it’s going to get to know you and it’s going to tell you what you want. If you want diversity of thought, talk to a human. Conflict is great. I’m excited for the next generation of conflict that’s going to happen in the workforce because everyone is really used to getting the yes. That’s how we’re setting ourselves up. But the real value comes from diversity of thought and figuring out how to have synergy with two people that have different thoughts and create something better than themselves.”

Nystrom said that as the use of AI becomes more prevalent, it will be critical to maintain human interaction.

“I love human interaction and AI is not a replacement for it. It’s important to keep connecting with individuals inside work, outside of work. Have AI tools to boost that productivity but don’t give up your co-workers.”

Barclay talked about increasing talk about regulation of AI, saying AI is not an excuse for bad behavior.

“It suits you to use common sense in a lot of cases. From a regulatory standpoint, AI is not an excuse for bad behavior. If there is behavior or outcomes that are against the law, let’s say discrimination as an example, or price-fixing among competitors, you can’t say AI told me to do this so that’s why I did it, so now it’s OK. I think that what a lot of regulation is trying to address is response use of AI and remain focused on outcomes. There’s a whole new area now around ethical use of AI and the degree to which you need a human in the loop. If it’s a low-risk situation. If you’re in an unregulated industry and you give a customer a wrong answer, it’s a really easy thing to fix. You can create a risk-reward matrix around it. And then there are things that are going to be really high risk, discrimination based on age or gender, price-fixing over competitors. Those are really high-risk activities, and I would always keep a human in the loop. When I think of regulations, I think of that as just another rule set, another thing you can feed to it as context and be like, don’t do these things.”