AI Gone Wrong Part 1

You're facing a key deadline and ask ChatGPT for help. You're in a hurry, what could go wrong?
The Mobile Experience
Generative AI tools are becoming an increasingly important part of our everyday workflow. They can help us write better, faster, and more efficiently. But what happens when they go wrong? In this series, we'll explore some of the ways that AI can go wrong and one very important reminder.
Imagine someone is super stressed out that your mobile experience is suffering, and they task you with the job of writing a summary of how to improve the mobile experience.
You're pressed for time, so you ask your friend ChatGPT about that mobile experience. The answer looks great and you forward the results to your colleague as your own original work.
You don't have time to read the results, but in your mind, you expect ChatGPT to return something somewhat useful like this:
Everything seems fine, until someone reads the actual results, which are not quite what you expected:
What Happened?
Earlier in the day, you had been chatting with ChatGPT about your forthcoming baby, and ChatGPT was still in baby mode when you asked about the mobile experience. So it ended up describing this:
What's the Lesson?
Verify All AI Results. Generative AI can help you get work done faster, but you should thoroughly review and edit all results to make sure they're fully accurate. And you should probably use the results as background material and not as the final work you give to a colleague.
Try Living Spec's AI Features Today and Share Your Feedback
We invite you to explore these features within Living Spec and provide us with your valuable feedback and suggestions. Our primary focus with AI is helping you be more productive and efficient, so we started with speech to text, dictation, sentence auto-completion, and general chat assistance. We're excited to see how you use these features and what you'd like to see next.