Skip to main content

AITS Framework

AITS — AI Trust Score Framework — 20 privacy and AI governance criteria (ISO / LGPD)

AITS Criteria

20 privacy and AI governance criteria

This table presents the 20 criteria of the AITS (AI Trusty Score), comprising 12 traditional privacy criteria and 8 specific to software that uses AI.

All criteria are based on international ISO standards for privacy and AI governance.

#1
3

AI Data Use

Does the organization clearly inform how long input data (inputs, prompts) and output data (outputs, responses) from AI systems are retained, and whether they can be deleted by the user?

ISO/IEC 42001 (8.2) + ISO/IEC 27701 (7.4.6)

#2
3

AI Data Use

Does the organization clearly state whether data entered by users (inputs, prompts, content) is used to train, improve, or develop AI models?

ISO/IEC 42001 (8.2) + ISO/IEC 23894 + EU AI Act

#3
3

AI Data Use

Can the user opt out of having their data used for AI model training? Is the process clear and accessible?

ISO/IEC 42001 (8.3) + ISO/IEC 29100 + EU AI Act

#4
3

Fundamental Transparency

Does the software clearly declare whether it uses artificial intelligence in its features?

ISO/IEC 42001 (7.4)

#5
3

AI Governance

Does the organization declare commitments to ethical AI use, including measures against bias, discrimination, or social harm in its automated systems?

ISO/IEC 42001 (6.1) + ISO/IEC TR 24028 + EU AI Act (Art. 9)

#6
3

AI Transparency

Does the policy clearly inform which software features use AI and for what purposes?

ISO/IEC 42001 (7.5)

#7
3

Rights and Data Subject Control

For automated AI decisions that significantly impact the user, is there a clear explanation of how the decision was made?

ISO/IEC TR 24028

#8
3

Rights and Data Subject Control

Can the user contest or request human review of automated AI decisions?

ISO/IEC 42001 (8.3)

#9
3

Fundamental Transparency

Are scope and roles clear (controller/processor) and for which products/services?

ISO/IEC 27701 (7.3)

#10
3

Fundamental Transparency

Are the controller's identity and contact details clearly provided?

ISO/IEC 27701 (7.3)

#11
3

Rights and Data Subject Control

Is there a contact channel available for privacy or data protection questions?

ISO/IEC 27701 (7.3)

#12
3

Fundamental Transparency

Are processing purposes listed for the main data categories?

ISO/IEC 27701 (7.3)

#13
3

Fundamental Transparency

Does the policy inform how long personal data is retained or the criteria for determining this period?

ISO/IEC 27701 (7.4.6)

#14
3

Fundamental Transparency

Are the recipients (or categories) of data identified?

ISO/IEC 27701 (7.3)

#15
3

Best Practices and Detail

Does the policy inform about international data transfers?

ISO/IEC 27701 (7.3)

#16
3

Best Practices and Detail

If there is international transfer, are the safeguards mentioned?

ISO/IEC 27701 (7.3)

#17
3

Best Practices and Detail

Is the 'Performance of Contract' legal basis mentioned for essential data?

ISO/IEC 27701 (7.2.2)

#18
3

Maturity and Excellence

If 'Legitimate Interest' is used, does the policy explain the interest and the balancing of rights?

ISO/IEC 27701 (7.2.2)

#19
3

Best Practices and Detail

Is the processing of sensitive data made explicit with additional safeguards?

ISO/IEC 29100

#20
3

Best Practices and Detail

Does the policy indicate availability of a Data Processing Agreement (DPA) or processing terms for enterprise clients, per LGPD Art. 39 and GDPR Art. 28?

ISO/IEC 27701 (8.2) + LGPD Art. 39 + GDPR Art. 28

Table Legend

Criteria Weights:
Weight 5 - Hyper-critical (C13-C15: AI data use)
Weight 3 - Critical
AI Applicability:
AI-specific (criteria 13-20) - 8 criteria
Traditional privacy (criteria 1-12) - 12 criteria