Keep Your Thinking Cap On: 5 AI Usage Guidelines

As someone who graduated with a humanities degree, it feels ironic informing others how to use an inhuman tool.

Yes, I was one of those people who questioned if my career would be in jeopardy from Artificial Intelligence. I was first exposed to ChatGPT in college, and once I saw how it rapidly spat out responses, I immediately felt intimidated. Would studying writing be worth it anymore? What about other possibly affected jobs?

The age of Artificial Intelligence is here — and growing rapidly. Some of you might be like me, harboring a fear of technological replacements and loss of humanness. Or, some of you may be thinking, “Hallelujah! I don’t have to spend an hour thinking of five variations for an email subject line.”

Regardless of your thoughts, the worst thing we could do is pretend it’s not integrating within every part of our workplace. What we can do, however, is discuss how to harness AI for good and still use our own critical thinking skills.

What is Artificial Intelligence?

Artificial Intelligence is a set of technologies that allow “computers and machines to reason, learn, and act in a way that would require human intelligence.” This technology cannot think on its own, and everything it “knows” is derived from human data.

A simplified example of AI is a computer chess game. The more often the computer interacts with the user, the better chess moves the computer develops. Translation: The more data and feedback we provide AI, the more optimized results we receive.

Here are some popular AI platforms and their common uses:

ChatGPT - language model: content generation, code generation

Adobe Firefly - creative model: text-to-image, text effects, text-to-video

Claude - language model: data analysis, summarizing text, writing & debugging code

While AI is certainly fascinating and helps make our jobs easier, it is not perfect. Humans are flawed, and so are the tools we create. Therefore, how do we use a human-thought technology that contains human-like imperfections while still holding onto our own critical thinking caps?

1.    Use AI as an assistant only

AI is a resource, not an expert. When creating with AI, you are the driver and the machine is the passenger. AI can tell you the fastest route or make suggested stops along the way, but ultimately, you decide which direction to take.

As an assistant…

-       AI can help brainstorm ideas or edit final drafts

-       AI can condense written work or create artistic variations

-       AI should not be your only source for research

-       AI should not be the main contributor in your projects… it should be you!

Just like a Wikipedia page can be a resource or guide, it should not be your final destination. You are the owner of your projects.

If your project fails from false information, data bias, or another aspect, the responsibility still falls on you, not the AI tool.

2.    Examine and refine your results

I highly advise to never directly copy anything from AI. Always put in your human touch by examining and refining results.

Think of it this way: If AI is created by other humans using their data, knowledge, and work, you are still unintentionally taking another human’s creation and passing it off as your own.

Most AI results are not going to be perfect after the first search. Unless you’re experienced in writing AI prompts or working with a near-human AI algorithm, you’ll still need to make modifications at some point in your project process.

A human touch can be…

-       Incorporating language unique to you and your business

-       Manually changing graphics to match a certain aesthetic

-       Adding social context to a data analysis

3.    Identify bias

Since humans have biases, AI does too. Healthcare AI is a great example of this.

AI algorithms for computer-aided diagnosis symptoms (CAD) have been found to return lower accuracy rates for black patients than white patients. Underrepresented data, just like underrepresentation in real life, can skew these algorithms and their results.

Just because AI is supposed to increase efficiency does not mean it’s correct. Check your results for bias, and if you’re able, suggest modifications to your company’s algorithms.

To identify bias, analyze from a qualitative lens. Ask…

-       Are certain groups over or underrepresented in my results?

-       Are social, cultural, or environmental factors considered?

4.    Consult the team

Before deep-diving into a project, ask your team or supervisor for their AI expectations and limits. If your company doesn’t include guidelines, now more than ever is the time to define those.

At some high schools and universities, students are placed on academic probation for using the AI tool Grammarly on their essays. While the students say Grammarly was only for editing, the TurnItIn algorithm (used for detecting bot-written content), told teachers and professors otherwise.

Most of us are familiar with the phrase “academic dishonesty,” but a similar “career dishonesty” can occur if these guidelines are not provided in the first place.

AI-limit conversations and policies are growing within education, but other companies and industries still need stated policies before any harm or discrediting occurs, not after.

I only have one guideline for rule #4…

-       Always ask if you’re unsure

5.    Listen to your gut

If it feels wrong, it probably is wrong. If it feels right, it probably is right.

A key aspect that distinguishes humans from human-thought technology is that we have a moral conscience. AI will never have independent morals, and even so, they will always be programmed by other people.

Before sharing your AI creations, data patterns, or suggestions, listen to your human self first.

Conclusion: Your humanness is still needed

As AI becomes more prevalent in our lives, now is not the time to toss out our own critical thinking caps. We should…

1. Use AI as an assistant only

2. Examine and refine our results

3. Identify bias

4. Consult the team

5. Listen to our gut

Automation can increase efficiency, but what is efficiency that does not consider the big picture of the consumers you’re serving? When we combine our talents with AI, we have the opportunity to create with another perspective in mind — or even call it out.