Top 5 prompt engineering techniques for a sports journalist to get the most out of generative AI

By Leo Knight

While you may not use AI to completely draft your stories, there are a number of creative and smart ways to improve your ideas for your audience. Using AI platforms and chatbots can reduce the users critical thinking, which overall is not what you want when creating a well informed body of work, therefore, be mindful of bias, hallucinations, Information security and data protection.

Prompt engineering is a technique used on generative AI platforms/LLM (Large language Models) to enhance the response obtained by the user, it can allow for more fruitful explanations.

1. Chain of thought Prompting- What is chain of thought (CoT) prompting? | IBM  This is a way of gathering information from the AI on how and why they came to the response they provided to you. for example you could ask "Give me the top five Manchester United players of all time" for an opinion piece or listicle you are writing, but to gather a greater understanding you could ask AI to explain why and how they came to the conclusions they responded to you with. It allows for more accountability and reduces the risk of false facts that you can check after the fact. Additionally it makes the response more logical and easier to interpret.

   
   

2. Contextual Prompting- The Role of Context in Prompt Engineering: A Comprehensive Overview - Media & Technology Group, LLCThis prompt technique is very important as the context you include in your prompt can hugely influence what response you get, providing context for AI is key as it generates more information for it to access and engage with. Context in it's most basic form is background information that generative AI platforms might not already have access to, like ChatGPT and many other platforms might only have a database as recent as 2023 or 2024 so they will need greater context provided by you in order to get a better response. Context can also be a number of different things like input context which can expand on things like what tone you want, the length of the response and specification of what audience you are catering for; for example FourFourTwo magazine or a BBC Sports audience. The more specific the prompt the better it will be.

   
   

3. Few shot prompting- Few shot prompting is a innovative way to improve the response you get by including examples in your prompt. for example you like an article you see online, you paste the good elements of that article as an example and use it to gain a better response from the generative AI platform. It also will give the platform more context for what the end goal of what you want in your response.

   
   

4. Role-Prompting- Role-Prompting: Does Adding Personas to Your Prompts Really Make a Difference?This technique is used a lot, sometimes you might not even know you are using it. This is when you give the AI platform human characteristics, for example, "you are an experienced sports journalist editor with years of experience in drafting and editing sports content stories" and then you follow up with the rest of your prompt. This in a way from now on will tailor the responses you get, it will definitely match the tone of the assigned persona you have given however it will not always be consistent and you will need to aware of that when using it persistently.

   
   

5. Failure allowance- 12 Prompt Engineering Techniques. Prompt Engineering can be described as… | by Cobus Greyling | Medium This key technique is when you allow the chatbot to admit when they don't know or are unsure about an issue. This will eliminate the platform from blindly making stuff up to appease you which will also rid of bias and transparency issues. For example you could input "Who won the Championship in the 2024/2025 season" or "Give me content ideas about Wimbledon" which current databases may not be updated on or well clued up on at all. It really is a solid fail safe in case AI tries to include any hallucinations at all.

   
   

   

   Cover image created by ChatGPT, Made 06/07/2025

   Remember, LLMs are still a working progress and very much a tool to help develop and further your work, as journalists we will always be held accountable and must maintain critical thinking throughout any work produced.