As Open AI released their prototype Chat GPT on November 30th, 2022 the hype and increased fascination with this chatbot continued to grow. Not only did it continue to grow, but it broke world records reaching 1 million users within the first week of its launch. To say the least, the world was taken by surprise as to the capabilities this chatbot possessed, and is truly looked at as one of a kind and a breakthrough within AI-generated chatbots.
But of course, with the hype came the growing concern as well. This concern was felt across the board by many people, and industries worrying if in the near future, they’d be replaced by this AI-generated technology. I myself was completely enthralled by the conversations, knowledge, and recommendations I was receiving from my interactions with Chat GPT. But the question still lingered in my mind: what are the limits, and boundaries I need to set up for myself to control and lead how I use Chat GPT? I’m sure others wondered the same.
So here are 10 commandments that were first mentioned by our CEO Eran Shlomo and should be adopted by all of us. Feel free to add your own commandments and share them with us.
The 10 Commandments of How to Use Chat GPT
I. Attribute the content to the author, by bot name, in the post.
To start, I’d like to take the opportunity here to address that you shouldn’t think of Chat GPT as your new intern who will do ALL your work for you, leaving you to sit back with your feet up. If you’re going to use Chat GPT as a reference then reference it just like you would with any other reference, or author.
II. Editing requires attribution.
This is a requirement that is often noted when working on a collaboration whereby multiple people contribute to a project over time. It’s only noble to give credit to an original creator in order for others to understand where the history of the work came from and who contributed to it.
III. Once edited, you own the result.
You start out by asking a question to Chat GPT and receive your answer. Think of that as your starting point. You’re brainstorming. Just like you’d do research for any piece of content and then turn it into your own. Think of it the same way here. Once, you’ve molded, modified, and manipulated the information and changed it into something you can call your own, and something you can stand by and defend, then you take ownership of that result.
And this should go without saying, but nevertheless, if you don’t agree with the information you rendered from Chat GPT, definitely do not under any circumstance post it. This also includes screening for harmful content.
It is also important to note here that while Chat GPT can generate coherent and grammatically correct text, it is not capable of picking up on nuances and subtleties of the context, therefore, nonsensical or even possibly unrealistic responses can be generated and therefore it’s important to always review and modify the text before posting it.
IV. Always use Chat GPT with a human-in-the-loop approach.
Chat GPT can assist in the creation of text, like blog posts, but human input is needed to fine-tune its performance. A human review is necessary to check the accuracy and relevance of the generated content. Human feedback allows the model to learn and improve over time. A human touch can also be added to the generated content to make it unique.
V. Fact checking is on you, reference fundamental facts in the post.
We all know that Chat GPT is known to have inaccuracies. Therefore, fact-checking is on your shoulders. For instance, if you ask Chat GPT to gather facts on a specific topic, find out where these facts originated from, verify them yourself, and don’t forget to note where the source comes from.
While Chat GPT is capable of generating a lot of accurate information, it does not fully comprehend the context in which you are using it, therefore it’s important to always verify the information and reference the sources.
VI. Remember personalization is on YOU.
Personalization here is two fold. Firstly, you have the ability to tailor and personalize the model’s output to suit your specific needs. It is important to remember that Chat GPT does not personalize the output automatically, this is something that you’ll be responsible for. In order to accomplish this you’ll to:
- Define your goal clearly: Having a clear goal in mind will help you personalize the results.
- Use the right prompt: Be specific in what you are asking and it’s all in how you word the prompt.
- You’ll need to fine-tune the model: When you receive an output, analyze it and if it is not what you were expecting, try to redefine or explain explicitly what you expect.
- Keep in mind the context: This will help you better fine tune the end result.
- Make it your own: Chat GPT is not a replacement for human creativity. Make the content unique by adding your own personal touch, and tailored to your specific needs and vision.
VII. Train Chat GPT for high-quality content, don’t expect it to work the other way around.
If you are expecting high quality content, it needs to come from you feeding information into Chat GPT, you can’t expect it to work the other way around. Get to know Chat GPT, yes, it is indeed like a relationship, invest in it. Teach Chat GPT what is correct and not correct. Don’t expect a language model to be the expert on a product, a concept, etc, you need to verify the expertise.
VIII. Creativity is on YOU not Chat GPT.
If you ask Chat GPT if it has the ability to be creative this is the response you get:
“As a machine learning model, I do not have the ability to be creative in the way that humans do. I am trained on a large dataset of text and can generate responses based on patterns to prompts, but it is not coming from an innate ability to be creative, but rather my ability to analyze patterns and make predictions based on what it has learned.”
The bottom line is the responsibility of creativity lies solely upon you, the creator of the content piece.
IX. Bias and discrimination are on you, even if it is automatically generated.
It is crucial to point out that Chat GPT is trained on a dataset that reflects the biases and stereotypes that exist in the data. Therefore, the responses generated by that model may perpetuate these biases and stereotypes even without being aware of them. It is important to be aware of this limitation and take steps to mitigate any potential bias or discrimination in the responses generated by the model. Again, the creators are responsible for the outcome.
X. Think outside of the box.
Always push forward and think of how you can use Chat GPT to improve your productivity, your angle, your position amongst competitors, etc. This is a tool that has no bounds for using your creativity so apply it to multiple areas. Some areas people are using it for include: creating emails, creating content, summarizing documents/text, translation, reviewing code, using email plug-ins for automated responses, get feedback on a piece of content. The possibilities are truly endless. Don’t be afraid to sit down and take the time to try out what works and what doesn’t work for you. Read articles about how others are using it, and use this as an opportunity for testing what works for you and your needs.
Success is In the Eye of the Beholder
When it comes to using Chat GPT success will vary for each individual or organization. Utilizing an AI-generated chatbot like Chat GPT is a powerful tool that has the potential to bear fruitful and positive outcomes, however, it should not be used to replace human judgment, creativity, or the essential role a human plays in creating content.
It is crucial to keep in mind the importance of attribution, fact-checking, and personalization when using Chat GPT. The goal of using Chat GPT should be to improve your own work, but not to completely automate it. Achieving success with Chat GPT will vary according to the task or scenario, and it should be noted that it may not always be the most suitable solution or provide the best outcome. Rather, you should use Chat GPT as a tool that can help you improve your productivity, efficiency, and even creativity if used with a solid set of rules or in this case commandments.
I encourage you to share your own experiences and suggestions for using Chat GPT effectively, and to continue to learn about the model and its immense capabilities, just remember the human-in-the-loop is crucial!