In recent weeks Social AF has been involved in an increasing number of conversations around charities using AI to help with moderation. These conversations have focused on one of two things:

1) Our opinion on how AI could add value to charities
2) If we are terrified AI is going to put us out of a job! 

As an agency that champions the need for real, human interactions at every turn it should come as no surprise that we have reservations about charities exploring AI as a potential solution for charities looking to improve social media moderation.

It’s undeniable that there’s somewhat of a crisis in the third sector. The demands and pressures on staff can be extreme and the volume of activity on social media is ever increasing. It makes sense that charities take the time to consider how investing in AI could help solve some problems.

The AI experiment!

We decided to conduct an experiment and ask AI to suggest some replies to posts found within our Facebook challenge groups. We then compared their suggestions against those produced by our expert moderators. Here are three examples:  

Post: I needed a challenge to get motivated. Doing this for a friend who is battling cancer.

Post: I’m going to drop out! I’m not getting any sponsorship at all!

Post: I was ahead of the miles but with it being half term I’m not sure how I’m going to manage with a lazy 4-year-old who isn’t keen on walking and I’m a single mum! We walked around the woods today and made it fun but she just about managed 2 miles and moaned the last one!

When you read the answers it’s easy to think they’re not bad replies. They’re warm, and positive and acknowledge the content of the post. It also took roughly 15 seconds to copy the supporter’s post, open the chat and ask for the reply, so you can understand how on face-value people think AI is the answer to all their prayers. 

But how do those answers compare to Social AF’s replies?

Post: I needed a challenge to get motivated. Doing this for a friend who is battling cancer.

Post: I’m going to drop out! I’m not getting any sponsorship at all!

Post: I was ahead of the miles but with it being half term I’m not sure how I’m going to manage with a lazy 4-year-old who isn’t keen on walking and I’m a single mum! We walked around the woods today and made it fun but she just about managed 2 miles and moaned the last one!

As you can see, the replies are worlds apart when it comes to quality and highlight the inadequacies of AI when it comes to charity moderation. 

Here are the things AI missed but expert social media moderation picks up on.

1) Acknowledging where the supporter is in their sign-up journey.

We add individual registration pushes or references to fundraising pages because we take the time to check. You’d have to spend an awful lot of money on a custom system, and then continue to invest in it, to do the same thing with AI.  

2) Referencing the charity/cause in a meaningful way. 

Again, we inject cause-led language and charity references to ensure the supporter knows the difference they are making. Whilst you can achieve this with some programs, in our opinion it looks forced and automated. By having a human crafting the response you can add references in a more meaningful way. 

3) Inclusion of appropriate questions and prompts

We love seeing conversations and often will ask questions to ensure people remain engaged. For AI, their MO is problem-solving so, unless asked, what they publish will just address one problem. It won’t look at the wider context, or spot the trends of multiple posts. One of our important roles is safeguarding. We look at the entirety of a supporter’s interaction. This means we’re better placed to include signposting to services when we spot something which is a cause for concern, something AI may well miss when responding in isolation. 

4) AI lacks true charity magic

As proud fundraisers, we know that the greatest asset to any charity team is YOU. Don’t risk losing what makes our sector so incredibly wonderful, by investing in something which strips that all away. 

Here are just a couple of interactions from our expert social media moderators in recent weeks to showcase the difference some real-life human banter can make.  

In summary:

How does Social AF think AI can add value to charities? 

The best way we’ve seen it be used is idea generation, to provide a jumping-off point or to improve something you’ve written.

We all have days where we struggle to immediately think of the right thing to say. If that ever happens to you, pop the question into ChatGPT and ask for a few variations. If you’re pulling together a content plan and can’t think of what to ask a group of runners, ask for 10 suggestions to get you started. Or if you’re writing a funding bid and you think you haven’t quite addressed the question, ask the AI program to improve and see what it comes back with. Then it’s a case of stealing the best bits and making it your own. 

Are we terrified AI is going to put us out of a job?

Absolutely not! It’s amazing how many third-sector problems can be solved by increasing capacity and ensuring the emphasis is placed on relationships and genuine connections. 

Managing resources may well be a problem you need to solve, but communicating with supporters shouldn’t be seen in the same way. Cutting corners with stewardship isn’t a long-term solution and charity’s biggest assets will always be the real people in it.

If you’d like to discuss how to make your social media Real AF, get in touch! We’d love to chat.