The Boundaries of ChatGPT: What it Can and Can’t Do

If you’ve been even remotely tuned into the tech world over the past six months, you’ve likely been bombarded with discussions about generative AI and ChatGPT. Everyone is talking about them. As a business leader, entrepreneur, founder, or product manager, it’s crucial to sift through the overwhelming amount of information and understand what this AI technology is truly capable of and its limitations. This knowledge will empower you to make informed decisions safeguarding your valuable time and budget.

To fully grasp the potential and limitations of generative AI, it’s essential to have a basic understanding of its technical context. In a nutshell, AI operates by identifying and learning patterns from a given dataset. Using these patterns, it can make decisions or suggestions without explicit programming.

Take music applications, for example, that analyze our past listening preferences and recommend the next song for our morning commute. These recommendations are based on a predefined library of songs. However, with generative AI, the possibilities extend even further. It has the potential to create entirely new and unique songs on demand, tailored precisely to each individual’s past listening history. It’s like having a personalized composer at your fingertips.

ChatGPT is built upon a powerful language model known as GPT (Generative Pre-trained Transformer). However, what sets ChatGPT apart is its explicit focus on facilitating conversations. When we engage with the model, we type a question or a prompt, and ChatGPT generates a response by analyzing our input. It’s important to note that while ChatGPT excels at generating text, it lacks a true comprehension of the prompts we provide. It understands the words we use and then crafts a response that closely aligns with our input. Whether we inquire about space exploration or seek recommendations for grocery shopping, the model treats each prompt with equal impartiality.

Check the video below for a practical business example of using ChatGPT to craft an email to welcome a new colleague.

You might be thinking, “Hey, I can search on the web and find dozens of templates for whatever I need.” However, leveraging ChatGPT in this scenario offers distinct advantages over a standard internet search. The output generated by ChatGPT is presented in a ready-to-use format, saving you time and effort in refining the content. Moreover, under the current Terms of Service on OpenAI’s website, you retain intellectual rights to the content generated by ChatGPT. This gives you the confidence and freedom to utilize the output as you see fit.

Furthermore, ChatGPT allows for easy customization to align the content with the specific needs of your business. You can achieve this by simply modifying your prompt or even by uploading your own relevant data. This flexibility empowers you to tailor the generated content precisely to your unique requirements. In subsequent articles, we will explore practical demonstrations showcasing how you can leverage these customization capabilities effectively.

Maybe ChatGPT cannot — and probably shouldn’t — write scripts for the next Netflix blockbuster. Hopefully, it will not generate content that draws millions of followers on social media. But this is different from what most of us write content for. In most cases, we use text to provide information, resolve a question or, in general, “get stuff done.” This is functional writing. The underlying assumption of these tasks is that there is no single correct answer. Instead, the model operates on the premise that certain responses are more likely to be valid than others. Applications like ChatGPT can streamline and enhance our content creation process. It offers us a valuable tool to generate high-quality text and significantly boost our productivity.

It’s critical to recognize that the responsibility for the generated content ultimately rests with the individual or entity utilizing ChatGPT or similar models for business applications. With that in mind, there are several best practices to consider:

• Evaluate the business impact. ChatGPT’s strength is understanding language and generating words they read as if a real person wrote them. Most of the time, the content is sensible and accurate. Sometimes though, the model makes up information, a phenomenon known as “model hallucination.” Before using it, consider how critical a specific business application is and whether someone must confirm the content before exposing it to readers, users, or clients.
• Test, test, and then test some more. Get in the shoes of your users or customers and interact with the model as if you were a user or customer. This firsthand experience will help you better understand the capabilities and limitations of the model, enabling you to provide more accurate and valuable content. This will give you a great insight into common errors and limitations.
• Transparency is key. Be transparent with your users and customers about the use of AI-generated content. Clearly label any content generated by an AI system, especially when human review is not involved. This transparency helps manage expectations and builds trust with your audience.
• Consider data biases. While you cannot directly influence the training data used for the model, you must be mindful of potential biases related to your specific use case. Conduct stress tests to assess how the model handles different scenarios and ensure that its responses are unbiased and fair.
• Differentiate yourself. As access to ChatGPT becomes more widespread, you must differentiate your content from competitors. Simply using the generated content as-is may lead to mediocre results. Take the time to edit and personalize the model’s output, adding your unique touch to make it stand out. An alternative way to customize your content is to modify the prompts you use, which we will cover in a later article.

By following these best practices, you can harness the power of ChatGPT effectively and mitigate potential challenges, ensuring that the generated content aligns with your business goals while maintaining high quality and user experience standards.

Thank you for reading! Let me know in the comments your thoughts and questions.

If you enjoyed this article, please give it a round of applause by hitting the clap button generously! Your support and engagement are greatly appreciated!

To stay updated on future AI insights and discussions in the business realm, follow the “AI with Lina” publication on Medium. For video content, follow me on YouTube. You can also connect with me directly on LinkedIn.