Paul, 13 November 2023

Use AI so it doesn't use you

Using AI tools such as Chat GPT can help you streamline content production and reduce your team’s workload.

However, you need to be aware of certain risks and challenges when using AI.

By obeying these five golden rules, you can stay in control:

Combine AI tools with human expertise.
Always start with user needs.
Check quality and accuracy.
Check for bias and fairness.
Make sure content is legally and ethically correct.

This will mean you can:

create high-quality content
use AI (and not the other way around)

1. Combine AI tools with human expertise

Remember that AI tools are most useful when combined with human expertise.

Yes, AI tools can automate some parts of content management. However, your team’s creativity, strategic thinking and editorial judgment can’t be replaced.

Your team’s human skills are vital for high-quality content that engages your audience.

2. Always start with user needs

It can be tempting to dive straight in and begin prompting AI tools such as Chat GPT to get content leads. However, it’s vital to get a real understanding of your audience first.

Good content always starts with quality user data. Let this data be your guide when you start asking AI for leads.

3. Check quality and accuracy

Never assume that everything an AI tool gives you is totally reliable. Instead, use it as a writing assistant for further research.

For example, one of the biggest complaints about ChatGPT is that it’s not accurate. That’s because it doesn’t provide:

sources
footnotes
links to the information used in the ChatGPT answers

Use prompts to get reliable responses

When you use ChatGPT, clearly specify that you need reliable and accurate sources. While this won’t always work, it might remind the AI to give you more useful responses.

For example, you could say:

“Please provide me with reputable sources to support my argument on (whatever the topic is).”

You can also tell ChatGPT the kinds of sources you want. When you’re looking for academic articles or books you can mention this in your prompt. For example:

“Please recommend peer-reviewed journals that discuss (and here, repeat what you discussed earlier in your conversation).”

Make sure information is up to date

Another idea is to use sources that don’t have ‘link rot’ (otherwise known as links that no longer exist online at the URL that ChatGPT recognises).

However, be careful, as ChatGPT doesn’t know about things after 2021. For example, ChatGPT couldn’t tell me who won the 2022 World Cup.

You might be tempted to use a prompt like this: “Please provide me with sources published within the past five years.”

Instead, consider using a prompt such as: “Please provide sources published from 2019 to 2021.”

4. Check for bias and fairness

Bias (showing a preference for one thing instead of another) is also a big issue when dealing with AI-generated content.

AI models are based on a large amount of information. Some might contain unseen preferences.

Any biases in the data can result in biased content. Biased content will be unfair as it favours certain groups of people or stereotypes.

Thankfully, there are AI-powered tools available that use linguistic analysis to flag potential biases. For example, you could try IBM Watson’s AI Fairness 360, which is free.

You can also look for bias by:

doing regular audits and evaluation – for example, assign team members to check for examples of bias or prejudice
promoting diversity and inclusion (for example, recruiting people from diverse backgrounds to leadership roles)

5. Make sure content is legally and ethically correct

You’ll need to make sure content generated by AI tools doesn’t infringe copyright laws or violate privacy rights.

Review and verify the output of AI tools to avoid any potential legal or ethical issues.