AI and IA (Inherited Assumptions)

Every time you use AI to write copy, personalise an email or filter an audience, you are working with a system that carries inherited assumptions.

Most people don’t really consider that fact. They check whether the content performs. They rarely check who it performs for, and who it quietly leaves out.

PwC’s report on algorithmic bias makes for sobering reading (and might make you reach for a glass of wine).

According to the research, AI is already influencing who gets a job, who gets a loan and what medical treatment a patient receives.

The root cause isn’t rogue technology.

It’s people: the teams who build the models, choose the data and interpret the results. Bias goes in at the design stage, and AI scales it.

Content is no different.

If the tools you use have learned from skewed data, the audiences you reach, the language you default to and the assumptions baked into your messaging will reflect that. Quietly. At scale.

The fix isn’t to stop using AI.

It’s to interrogate what it produces. Who is this written for? Who does it exclude? Would someone outside your typical customer profile find this relevant or invisible?

Those are editorial questions. They always were.

Ready to Work Out Where AI Fits in Your Marketing?

If you’d like an honest conversation about where AI can help your marketing (and where it can’t), book a free 20-minute Connect Call. I’ll tell you what I think.