1. Be Skeptical

Remember that AI often makes mistakes, whether that be pulling data from thin air, errors referred to as “hallucinations”, or providing users with inaccurate information. In fact, a study conducted by researchers with the Colombia Journalism Review revealed that ChatGPT was “confidently wrong” 71% of the time when given a quote and asked to cite the article. Only rarely (3.5% of the time) did the AI admit to being unsure or unable to carry out the task, displaying an alarming amount of overconfidence. 

Being skeptical and double-checking facts with other sources is the best way for students to ensure they’re not being misled by a tool that is supposed to make them more productive.

2. Be Aware of Biases

Believe it or not, AI displays biases in a multitude of ways, such as interpreting information in a way that supports preconceived notions (confirmation bias), making decisions based on a limited and unrepresentative population (selection bias), and even referencing racial or societal stereotypes. According to a study published by the Association for Computational Linguistics, machine learning models are prone to making assumptions, such as occupation based on gender or skill sets based on race. This is inevitable to some degree, as all humans have biases, and AI is trained using datasets created by people.

Additionally, AI is known to amplify and agree with the biases of its user, often because they coincide with what it’s been trained on or because AI tends to be a “yes-man” to “flatter” its clients, even going so far as to apologize and alter their response if a user signals doubt towards the validity of it, even when the information was correct.

All of this is to say that AI isn’t only prone to misinformation when it comes to statistics and datasets. Therefore, being conscious of the potential limitations of the model you’re using is of utmost importance if you want to get the most out of your experience.

3. Use Your Mind First

It is becoming more common to use AI as a brainstorming and ideation tool, which can serve as a great starting point when you’re in a rut. However, artificial intelligence tends to prioritize quantity over quality, meaning it can produce a lot of ideas, but they will lack originality and focus on safe, conventional options. Consider these brainstorming tips to come up with ideas on your own, then use tools to develop them further:

• Come up with as many ideas as possible, no matter how crazy, mundane, unfeasible, etc, they may seem. Judgement can come afterwards!
• Consult your team, friends, or professionals you trust for perspective
• Use a visual method, such as mind maps, flow charts, or affinity mapping