Use AI. Don't use AI. Nobody says what that means at work.

At Sage, we’re encouraged to use AI.

That makes sense to me. Sage is a technology company. AI is becoming part of how modern businesses operate, and employees should be experimenting with these tools, learning where they help and becoming more productive with them.

I love that.

But then I open LinkedIn or read another technology story and encounter a completely different message.

Don’t use AI to write. Don’t rely on AI too much. AI is destroying creativity. AI is making us stupid. AI-generated work is somehow less legitimate.

And I keep wondering what the average employee is supposed to make of all this.

I spend a lot of time thinking about AI, experimenting with different models and working out where I want human judgement to sit. Most people don’t. They have a job to do.

If their employer is telling them to adopt AI, become more productive and find ways to incorporate it into their work, I don’t think we can then guilt-trip them for doing exactly that.

That doesn’t mean anything goes. People still need to take responsibility for what they produce. They need to check facts, protect confidential information, understand the work and know when human judgement matters.

But perhaps organisations have a responsibility here too. “Use AI” isn’t really enough guidance.

Where is AI encouraged? Where does human ownership matter? What are employees accountable for? What does good AI-assisted work actually look like?

I’ve written before about the gap this leaves — nobody designed the replacement for what production work used to teach, and it doesn’t have an owner because it was never anyone’s job to design.

I don’t have a neat answer. I just think the conversation needs to move beyond AI good, AI bad.

Because for millions of employees, AI isn’t some philosophical question about the future anymore. It’s becoming part of the job.

Update: I have now tried to turn this problem into practical advice: minimum viable AI for people who don’t want to use it.