Are You Prepared for an AI-Related Crisis?

I saw a post on Facebook this week from someone I am acquainted with. Not someone I know well at all, who has over a thousand "friends".

She was posting about the sadness of dropping her child off at university. A fair post - but I was deeply troubled by the images. There was a lovely picture of her child, wearing a hoody showing the sports team they play for. A picture of the university sign. I had the child’s full name. And the last image was the child hanging out of their new accommodation window, with a sign next to them giving the name of the accommodation and a range of 10 room numbers.

I found myself thinking: have we forgotten how much personal information we are giving away online, and who might be watching? This is more pertinent now that we're all giving our information to AI so freely.

I am in some charity and third sector Facebook groups, and a charity asked this week about moving away from using original pictures of children and uploading the images to AI to anonymise them.

I don't think I can stress enough how much of a bad idea this is. You'd almost be better uploading the picture of the actual child, with consent, than uploading that child's image to an AI server. That opens a whole new can of worms. Do your consent policies include uploading images to AI? Do you even have an AI policy in place?

I've seen numerous stories of people inputting sensitive data into AI (GDPR risk anyone?), finances being uploaded and presentations being created without double checking the numbers, resulting in them being wrong (costing huge amounts of money).

There was a news story recently where an MP used AI to generate a case study about a fictional nurse who was struggling to find affordable housing. Why do we need to create an AI case study when there are real people who are actually experiencing these issues?

All of these are a breach of trust and will damage your reputation, and ultimately that's what organisations rely on from the public. Organisations spend huge amounts of time building trust with the people they serve. But AI can create a crisis very quickly if we haven’t thought carefully about how we’re using it.

Some organisations are prepared for the crises they can see coming: financial problems, natural disasters, rogue volunteers, serious safeguarding incidents. But what about the crises you can’t see coming? Do you have the policies and crisis comms processes in place to stop your AI usage becoming a crisis?

It's not something I had thought about when I started my consultancy three years ago, and even before then when I was a journalist, this wasn’t part of the crisis landscape I was dealing with. I’m having to learn very quickly, so I can help you when the inevitable AI related crisis comes along.

If it's not something you've thought about and you're not sure where to start, I’ll be exploring this further at the CharityComms AI Conference in November.

But in the meantime, I ask again - are you prepared?