The right way to use AI lies between ban and free rein

This week's "How to" stories paint a clear pattern: the more artificial intelligence can do on its own, the more important it becomes how we humans set the boundaries around it.

It's clearest in schools. A two-year trial shows that students who aren't allowed to use AI at all actually do worse than both students with free access and students who have received structured training in using the tools. The conclusion is simple: bans don't solve the problem, but clear guidance does.

The same picture is emerging in the workplace. A new survey shows that half of Danish companies have lost track of how employees use AI in their daily work. The solution isn't to ban the tools but to set up simple guidelines, without stifling the urge to experiment.

AI takes more steps on its own

At the same time, AI itself is getting better at taking over several steps. Chatbots like ChatGPT can increasingly work out several steps on their own from one short message, instead of you having to guide them all the way. Google is going further and testing an AI agent, meaning a program that carries out tasks on its own, that is meant to coordinate the whole family's calendar, shopping and meal plans. OpenAI has meanwhile launched a new tool that lets companies query their own data in plain language and get answers as clear overviews.

The more we put in the hands of AI, the more pressing the question of what actually happens to what we write becomes. OpenAI has put hundreds of contract workers to read real ChatGPT conversations as part of an internal project to make the chatbot less flattering toward users. It's a reminder that your conversations aren't necessarily seen only by you.

What happens to your data

Finally, there's the more personal part of the equation: how to use AI yourself without letting it take over. A well-known tech figure uses language models only to check his texts, never to write them, and sticks to two simple rules to preserve his own voice.

Together, this week's stories point in the same direction: AI is getting better and more independent, but the payoff only comes if humans set clear boundaries for its use, at home, at work and at school.

Watch whether your workplace has clear guidelines for AI, and think about what you share in your conversations. If you want to give AI more responsibility, you can start small, for example with a single task at home, and at the same time spend a few minutes considering how you hold on to your own voice once the text is written.