Operational metadata is not used for marketing without consent and not used to train AI
For operational and product metadata, JetBrains says that unless the user gives consent it does not use this data for any marketing activities, and that it does not train its AI on this data. The license table repeats the marketing point for products under the JetBrains AI Terms of Service: for marketing purposes JetBrains only starts collecting analytics with the user's consent.
Supporting quotations
Unless you provide us with your consent, we do not use this data for any marketing activities.
We also do not train our AI on this data.
For marketing purposes, we only start collecting analytics with your consent.
Qualifications & exceptions
The marketing exclusion lifts if the user provides consent.
The AI-training exclusion is stated with no consent exception.
Limited to the operational and product metadata described in item b).
Not stated here
Section 2 lists marketing and training "our own AI models" among the purposes of collection; the notice does not map those purposes to each category beyond these exclusions.
Behavioral data is not used to train code or text generation models and excludes personal data and source code
JetBrains says behavioral data is not used for training ML models that generate code or text, or for other types of data from which outputs can be extracted. It also says behavioral data does not include any personally identifiable data or any source code files or fragments from the user project.
Supporting quotations
It is not used for training ML models that generate code or text, or for other types of data from which outputs can be extracted.
Behavioral data does not include any personally identifiable data or any source code files or fragments from the user project.
Qualifications & exceptions
Stated for behavioral data (item d) only.
Not stated here
The phrase "or for other types of data from which outputs can be extracted" is not defined, and its grammar does not make clear whether it adds a second kind of excluded model training or a second kind of excluded data.
Detailed code-related data can be used for product improvement and to train AI models
JetBrains says detailed code-related data can be used for product improvement and for training AI models. This differs from operational and product metadata and event-based data, for which JetBrains says it does not train its AI.
Supporting quotations
Detailed code-related data can be used for product improvement and training AI models.
Qualifications & exceptions
"Can be used" is a permission; the notice does not say every record is used for each purpose.
Not stated here
Section 1 f) does not say whose AI models are trained; section 2 names training "our own AI models" as a purpose.
JetBrains lists five purposes for which it collects only the data necessary
JetBrains says its primary goal is to provide tailor-made, data-based tools and that it collects only the data necessary for five purposes: product analytics, model evaluation, security and audibility, training its own AI models, and marketing.
Supporting quotations
Our primary goal is to provide you with tailor-made, data-based tools.
we collect only the data necessary for the following purposes: product analytics, model evaluation, security and audibility, training our own AI models, and marketing.
Qualifications & exceptions
Limited to "only the data necessary" for the listed purposes, as JetBrains describes it.
Section 1 narrows some categories: operational and product metadata and event-based data are not used to train AI, and operational and product metadata is not used for marketing without consent.
Not stated here
The notice does not map each purpose to each data category.
"Security and audibility" appears as written in the source; the notice does not explain "audibility".