Prompt Engineering Basics for Everyone

The AI revolution is here but it’s not yet the time to carry arms against bots. This time we’re controlling the bots. 

Artificial Intelligence has begun changing many aspects of our lives, especially, how we work and learn. My friend who studied Microbiology, is using ChatGPT to learn Statistics. Many content creators are already using AI tools to create content in batches and schedule them. I used Bing AI to generate the image below, with the prompt: ‘Create an image of a dark-skinned woman teaching a robot how to dance’.

What is prompt engineering?

According to Prompt Engineering Guide, prompt engineering is a relatively new discipline for developing and optimising prompts to efficiently use language models (LMs) for a wide variety of applications and research topics. Prompt engineering skills help to better understand the capabilities and limitations of large language models (LLMs) that power generative AI chatbots like ChatGPT, Bing AI, and Google’s Bard. 

Although prompt engineering is more than just creating and writing prompts, it involves carefully choosing the words and phrases we use to communicate with AI tools. We want to make sure that our instructions are easy to understand and give concise information. It’s a bit like writing a recipe. If you miss a step or use confusing language, the computer program might get confused and not produce the result you want.

Prompting or prompt engineering is how you can use artificial intelligence to accomplish your goals without having to learn complex coding or technical jargon. It helps you communicate effectively with AI to accomplish different tasks such as automating emails, creating images, creative writing, and brainstorming.

Why prompt engineering is a valuable skill

Prompt engineering is a highly valuable skill for individuals seeking to harness the power of AI to accomplish their personal or professional objectives. Here are some of the benefits of learning prompt engineering:

• It is easy and fun: You don’t need any technical background or learn any programming languages to become a good prompt engineer. You just need to know how to write or speak in a natural language. You can also experiment with different prompts and see what the AI produces.

• It is versatile and flexible: You can use prompt engineering for a variety of tasks, depending on your needs and interests. You can also customize your prompts to suit your preferences and style.

• It is powerful and effective: Prompt engineering can help you leverage the capabilities and potential of AI to enhance your work and life. You can use prompt engineering to create high-quality content, find relevant information, generate new ideas, and solve complex problems. Prompt engineering can also save you time and effort by automating some of the tasks you would otherwise have to do manually.

• As a relatively new discipline, there is a demand for people with prompt engineering skills. According to Time Magazine, salaries can go as high as $335k, without requiring a Computer degree.

How to use prompt engineering

To use prompt engineering, you need to follow these steps:

• Choose your preferred AI tool.
• Define the goal or task that you want the AI to perform.
• Craft a prompt that describes your goal or task in natural language.
• Give the prompt to the AI tool and wait for the output.
• Evaluate the output and refine your prompt if needed.

Although AI chatbots can understand natural language, you need to pre-prompt them to get better results. Pre-prompting is a way of familiarising the AI tool with the concept and scope of your work. An example will look like this:

“I am a student and I need your help writing a free verse poem on love, life and fashion. This poem should be relatable to everyone. Do not start writing till you confirm you have understood me. Okay?”

Then, the chatbot, in this case, ChatGPT, will request for more prompts, and so on.

Pre-prompting ensures that the output you get is more tailored to your desired result than a general output.

Here are some examples of how you can use prompt engineering for various tasks:

• Image creation: To create an image using AI, you can write a prompt like “Create an image of a Burmese cat wearing a hat” or “Create an image of a red sunset over a hill with some trees”. The image below was created by Bing AI powered by DALL-E. It can be further refined by adding more details in your prompt.

• Content generation and brainstorming: To get AI to write content for you or suggest content ideas, you can give it a prompt like “Write a blog post about prompt engineering for everyone” or “Write a product review for ChatGPT”. It will generate a basic article. To create a more refined content, you have to prime the chatbot. Start with a prompt like: ” I am a Literature student looking to write an article on How To Become a Good Movie Critic. Can you generate content ideas on the topic? “

• Voice command: Even before the current AI wave, we have been using voice commands to get AI like Alexa, Siri, and Google Assistant to do tasks for us.

• Reasoning: If you want an AI to solve a problem for you using logic and reasoning, you can give it a prompt like “Solve this math problem: 2x + 3 = 11” or “Answer this question: Who is the president of the United States?”.

AI is a powerful tool, but it is not perfect. Always double-check the results before using them.

How to learn prompt engineering

Many online resources are available to help you get started on learning prompt engineering. One of them is Learn Prompting, a free and open-source curriculum that teaches you how to use ChatGPT and other AI tools to accomplish your goals using prompt engineering.

You might not need to become a professional prompt engineer but if you want to optimize the human-AI collaboration, look into short courses on the topic, especially if you frequently use AI tools in your daily tasks. There are loads of free ones like this one on LinkedIn Learning and this ChatGPT course on Codecademy.

There are also many YouTube tutorials on AI prompting but this one by MIT AI ML Club comes highly recommended.

If you prefer listening to podcasts, check out The AI Podcast by NVIDIA.

Following the release of ChatGPT, many SaaS companies have rolled out AI enhancement features for their products and even their own conversational AI models.  AI chatbots like Google’s Bard and Bing’s AI are a few that currently have features like voice-to-text and text-to-image. More enhanced AI tools will be released before the year runs out.

Anyone can learn and use prompting to communicate with AI effectively. By mastering prompting, you can unlock the power of AI and get better than average results with ease and fun.

These are still the early days of Generative AI and we are yet to grasp its full effect, but those that adapt quickly to these changes will be the better for it. So what are you waiting for? Start learning prompt engineering today!