A BEGINNER’S GUIDE TO USING AI

YOUR FIRST 10 HOURS

WHENEVER I speak to a group about artificial intelligence, whether they be business executives or students, I ask them about their AI use, and I am always surprised that only a handful of individuals have spent the time needed, with the right AI systems, to give them a sense of how useful this tool can be. Most seem to believe that, if they wait, someone will give them exact instructions about how AI can help them. For them, I have good news and bad news. The bad news is that there is no instruction manual out there that will tell you how to best apply AI to your specific job or interests. That’s a shame, because AI is good at some tasks and bad at others in ways that are difficult to predict if you haven’t used it a lot. The good news is that, by using it for only 10 hours, you can figure out the best way for you to move ahead. Ten hours may seem like a lot, but just think about how many hours a day you spend on social media or watching television. Getting started I want to give you some tips on how to spend those 10 hours most effectively. First, you need to choose an AI tool. Pick one of the smartest BY ETHAN MOLLICK systems available, allowing you to gain the fullest understanding of the possibilities. That narrows the field down to three: OpenAI’s GPT-4, Google’s Gemini Advanced or Anthropic’s Claude 3 Opus. Once you have a model picked out, try it out on the things you do every day. Think of it as an intern. See how it can help you compose email, write a report, summarize and analyse data. But don’t stop with relatively mundane tasks—ask it to get creative. Has it write code to help you be more efficient, create images, produce videos? Generate ideas. The more you ask it to do, the more you’ll learn about its capabilities and the better you’ll understand its strengths and weaknesses. When interacting with AI, don’t get bogged down by the idea of crafting perfect prompts. Instead, focus on having a natural conversation and providing clear instructions. The best results come from a back-and-forth interaction, not a flawless first prompt. Break down complex tasks into step-by-step instructions. Provide examples of what you’re looking for—give it a chunk of your own writing, for instance, and tell it to write in that style. You can even assign the AI a persona to help it understand its role. You can tell it to “act as a marketer for snack foods” or “play the role of a stock-market analyst.” These commands won’t magically transform the AI into an expert, but it will shape the AI’s answer in ways that can be very useful. Expect mistakes, such as hallucinations—content with no basis in reality, like a made-up quote or a factual inaccuracy—or information that is outdated. Or it may simply fall short of what you were hoping for. When that happens, don’t give up. Tell it where it went wrong and keep asking for something closer to what you want. Often that means providing additional information in your prompts. There’s no end to how many re[1]visions you can request—unlike a human intern, it won’t get tired or discouraged. The next level After you have gained a little experience, focus on applying AI to areas where you have deep expertise. That way you can better understand how it can help you in your particular field and what its limitations are in that area. Again, think of it as an intern—but now as one you can push farther. Assign it tasks, provide guidance and give it feedback. This approach will help you identify where AI can add the most value to your work. Finally, it’s important to keep in mind that AI tools will become increasingly sophisticated, accurate and efficient. So think about not only what it can do for you now, but also about how it might be able to help you succeed in the future.

Ethan Mollick is an associate professor of management at the Wharton School of the University of Pennsylvania.