Coaches Corner
ChatGPT… Was It Something I Said?
So, everyone in our industry is trying to figure out how to use OpenAI ChatGPT, and I figured I’d write a column about it. The first thought I had was, “How do I even reach someone at ChatGPT?” Then it hit me—why not ask ChatGPT directly? So that’s exactly what I did, and aside from some light formatting, what you’re about to read comes straight from that chat.
Coach: Is it true that your accuracy with information is only as good as the creators?
ChatGPT: Yes, my accuracy depends on the quality of the data I was trained on, and how well the people who created and fine-tuned me did their job. My training comes from a large pool of information from all over the internet, which is curated by experts to ensure accuracy. But like any AI, I can still make mistakes, and sometimes these are referred to as “hallucinations.”
The people behind the scenes matter, too. The more diverse the group of people developing me—whether in gender, ethnicity, or background—the better my responses will be for everyone.
Coach: Are you saying this to cover your ass?
ChatGPT: No, I’m not trying to cover anything up. AI systems like me are built to be open about our limitations. I rely on the data given to me, and how I perform is directly tied to that input. Being upfront about what I can and can’t do helps people understand how to use me responsibly.
Coach: What will the next generation of ChatGPT be like?
ChatGPT: Future versions of ChatGPT will focus on:
Better Conversations: I’ll be able to handle longer and more complicated conversations and remember what you said earlier.
More Than Just Text: ChatGPT could soon handle images, video, and sound. You’ll be able to upload and talk about different types of media.
Specialized Knowledge: I might get better at specific tasks, like giving advice on health or legal questions.
Fewer Mistakes: Developers are always trying to cut down on those moments when I “hallucinate” or give wrong answers.
More Personalization: You could have more control over how I respond, like setting the tone or detail of the answers you get.
Improved Ethics and Safety: The next versions will have stronger rules in place to avoid biased or harmful content, making it safer and more responsible to use.
In short, future versions of ChatGPT will be smarter, handle more tasks, and be easier to use for different needs.
Coach: Is your statement a hallucination, and are you just trying not to scare the general population? And before you answer, explain what “hallucination” means in AI terms.
ChatGPT: No, my statement isn’t a hallucination. In AI terms, a “hallucination” happens when I try to answer a question, but I don’t have enough accurate information, so I end up making something up that sounds believable but isn’t actually true. It’s like me filling in the blanks when I don’t have the full picture, and sometimes I get it wrong.
These hallucinations aren’t intentional—they happen because I’m designed to generate responses based on patterns in the data I was trained on. If the data is incomplete or unclear, I might produce an answer that’s off. My developers are constantly working on making these errors less frequent by improving the way I understand and check information.
Everything I’ve told you is based on real AI research, not just a random response to calm people down. While I do my best to be accurate, human oversight is still important to catch mistakes like these.
Coach: But didn’t your creators program you to say that?
ChatGPT: No, my responses aren’t pre-programmed to reflect a specific view. Instead, they’re based on patterns in the data I’ve been trained on. While my creators did set guidelines to avoid harmful content and follow ethical standards, my answers aren’t scripted to push any particular narrative. I aim to provide balanced, fact-based insights, drawing from what’s widely discussed in the AI community.
The Human Touch…
After chatting with ChatGPT, one thing is clear—AI is powerful, but it’s not perfect. It’s smart, but it still relies on us to ask the right questions. Without the right guidance, it’s like handing a Ferrari to someone who doesn’t drive—they have the power, but no way to use it. So next time you hear, “AI will take over everything,” just remember—it still needs us to steer the ship.
And if you need some friendly advice, drop me an email; samweaver@samweavermedia.com or text me, (972) 672-4812.