Your data and AI: what each tool reads, and where it sends it.
For each AI in your software: what it reads in your company, where it sends it, whether it learns from it, and the protected version you may already have.
The question
“What happens to our data in these tools?” Each vendor has its own terms, and they change from one edition to another: the same feature can train its models on your content in one edition and not in another. Nobody has the time to read the terms of twenty pieces of software, or to follow their updates.
The answer: five steps
- A data fact sheet per tool and per edition: what it reads, where the data goes, whether it trains its models on it.
- The protected version you may already have: when a protected edition exists, the fact sheet says whether it is included in your licenses.
- Your rule next to the tool, in the employee portal.
- An alert when a vendor changes its terms.
- The setting that already exists: when the tool lets you turn off training or keep data in Europe, the fact sheet points it out; your administrator applies it.
The evidence
The free readout: its third count, read in your tools, says how many of your AIs train their models on your data or send it outside the EU.
What Ojja does not do
- Ojja reads registers and configurations, never your content.
- Ojja does not say whether a use is permitted: it gives the facts, and the points to review with your counsel.
Frequently asked questions
Is our data used to train the models?
It depends on the tool and its edition. Each tool's data fact sheet says so, with its source and its date.
Does our data leave the European Union?
The fact sheet says where the vendor hosts and where it transfers to, edition by edition.
Does Ojja read our content?
No. We read registers and configurations: licenses, settings, connected applications. Never an email, a document or a conversation.