Memory: What your AI knows about you

Many modern AI assistants can use information from previous conversations or your user profile: your profession, preferences, writing style or topics you frequently discuss.

That is convenient in everyday use because you do not have to start from scratch every time. But it also creates an effect that is easy to overlook.

The invisible filter

Stored context can change more than just tone. If an AI already knows something about you, that information can influence which aspects it selects, how strongly it weighs them and in what order it presents them.

• What comes first? Some arguments may be given more prominence than others.
• What moves into the background? Counterarguments or alternative perspectives may receive less attention.
• Which risks matter? The same situation may be framed differently depending on what the system already knows about the user.

Two people can therefore ask exactly the same question and receive answers based on the same facts – yet with different emphasis, creating a different overall impression.

Your leverage in everyday use

• The check: Occasionally look at what information your AI assistant has stored or uses for personalisation. Where and how this works differs between ChatGPT, Gemini, Claude and other systems.
• The neutrality test: For important decisions, it can be useful to ask the same question again in a minimally personalised or temporary chat and compare the two answers.

The key point: Memory does more than make AI convenient. It changes the context from which an answer is produced.