Wait, what do you mean? 

YES! NO!

You will not necessarily need to learn prompt engineering, even though it might be all that anyone is swearing by at the moment.

Let me de-confuse you.

Who even is a prompt engineer?

A prompt engineer is responsible for fine tuning and modifying existing AI Models to perform better. They fine-tune prompts through continuous and repeated modifications, data filtering, and pre-set rules.

Why should we care? 

The main interaction between humans and Artificial Intelligence bots like ChatGPT works in prompts.

We can speak to it in the same way we interact with each other. That’s the whole reason ChatGPT became so popular.

Knowing how to prompt correctly (i.e., how to properly speak to ChatGPT) can help increase the accuracy, coherence, and quality of the outputs AI generates.

But knowing how to design a good prompt does not equal prompt engineering.

So who actually needs Prompt Engineering?

It depends what you’re trying to do and whether you are using supervised AI or unsupervised AI.

You don’t need it when…

You are sitting there, typing the prompt into ChatGPT, you don’t need prompt engineering. You can go back and forth just like you would with an assistant. For most people, this is their use case.

When you are there to do quality control, to ensure the output is as you wanted it to be and go back to question it.

When you give feedback to revise / enhance an output.

Or when the answer you receive does not have to follow a specific format.

When you don’t need always to include the same information to prompt.

When you need a one-time output that doesn’t repeat.

Then, you do not need Prompt Engineering.

Just to be super clear, here are some practical examples:

• Checking Documents to review / summarise general understanding

• Brainstorming Ideas / Campaigns / High-level strategies

• Conduct Market Research

• Brainstorm New Product Ideas

• Check Grammar and Spelling

You do need it when…

When you (the human) are not there all the time to give feedback, revise outputs, control quality - yet you still need to ensure consistency and reliability of the agent.

Or when you need a very specific output generated. Or when you happen to use the same prompt many many times, and they have to be in line with certain data inputs.

Practical examples are:

• Automating customer support responses

• Generating routine reports

• Automating repetitive tasks

• Drafting consistent email outputs

Anything that absolutely requires consistency and reliability of the output.

So, to sum it up. 

In your day to day operations and when using ChatGPT for basic tasks, you probably won’t need prompt engineering.

However, when setting up automated processes that require consistent, high-quality outputs, it’s important to consider prompt engineering best practices.

Additionally, there are some effective prompting strategies you can use even for everyday tasks.

We’ll share some tips and tricks on how to structure your prompts to get the best responses from ChatGPT in one of the next issues.

In the meantime, if you are not a member of the AI Leadership Forum, join today and become part of the next cohort of applicants!

This Week in AI

• OpenAI is reportedly planning to develop its own AI Chip;

• LinkedIn is looking to implement AI Games to encourage platform use;

• The UN is set to open a forum to address AI Governance globally.