Journal / Opinion · 2 min read
What does a trustworthy internet look like?
Trust plays out in concrete moments: knowing which information is used, why, and what you can still decide.

Understanding what's happening
When a service suggests an action, you should be able to follow the reasoning behind it: what information was taken into account? What came from me, and what did the system infer? The answer to these questions makes it possible to correct the context rather than simply accept a suggestion.
Choosing before delegating
Preparing a message and sending it do not have the same effect. This difference must be visible in the interface. A good suggestion leaves time to review it and offers a clear choice at the moment of acting.
Staying in control
A trustworthy system is measured by how easily you can reconsider your choices. This is a design principle we apply to our thinking on GEN and the future uses of AI.