Have you ever found yourself asking, "How many 'R's are in the word 'strawberry'?" Probably not, but if you were to ask yourself this question, at first glance, it would seem like a straightforward task. You might even be tempted to quickly count and say, "Three, of course!" But when some people asked an AI language model like ChatGPT this same question, they received an unexpected answer: "Two." 🤔
Now, this might seem like a small mistake, but it actually highlights an important aspect of how we interact with AI systems—and why the way we ask questions matters more than we might think.
Get ahead of the AI curve with T-Tech's AI Academy. find out more here
The Miscommunication Behind the 'R'
So why did the AI get it wrong? The simple answer is that the model interpreted the question in a way that a human likely wouldn't. Instead of counting all instances of the letter 'r' in the word "strawberry," it provided the number of distinct letters. While this might sound like a minor issue, it illustrates a broader challenge: AI systems are only as effective as the prompts they’re given.
This isn't just a quirky anecdote about a letter—it's a lesson in how we should approach AI, especially as these systems become increasingly integrated into our lives.
Why the Way We Ask Questions Matters
When working with AI, the quality of the output is often directly related to the clarity of the input. Here are three key tips to consider when interacting with AI to ensure you get the most accurate results:
The way you frame a question or command can significantly impact the quality of the response. AI models, like ChatGPT or Microsoft Copilot excel when given specific, clear instructions. If you ask a vague or ambiguous question, the AI might interpret it in an unexpected way, leading to less relevant answers. For instance, instead of asking, "How many 'R's are in 'strawberry'?", you might ask, "Count all instances of the letter 'R' in the word 'strawberry.'"
No AI model is perfect. While they are powerful tools capable of processing vast amounts of information, they also have limitations. Understanding these limitations is crucial for effectively using AI. These models rely heavily on the context provided by the user. If the context is unclear, the results may not meet expectations. Recognising that AI models may not always interpret queries as intended can help users better manage their interactions and refine their questions.
Don't be discouraged if the first response you get isn't perfect. Iteration is a natural part of the process when working with AI. Sometimes, you need to refine and rephrase your question to get the desired result. This iterative process is not a sign of failure but a way to optimise interactions with AI and ensure you get the most accurate and relevant information.
Adapting to an AI-Driven World
As AI continues to evolve, so too must our approach to using it. Whether you're leveraging AI in business, research, or everyday tasks, good prompting is the foundation of effective AI use. The better we get at asking questions, the more useful and accurate the answers will be.
If you're interested in mastering the art of AI and unlocking its potential in your day-to-day work, consider enrolling in T-Tech's AI Academy. This 8-week AI mastery program is designed to empower accountants with the latest AI knowledge, helping them navigate the complexities of AI and harness its power effectively. For more information, visit our AI Academy home page
By understanding how to communicate with AI, we can ensure that we're not just getting answers, but the right answers. And sometimes, that starts with something as simple as counting the letters in a word like "strawberry."