Solutions
Automation for AI Risk Assessments
Our Process
Track model performance and reduce risk in three steps:
Evaluate
Evaluate models to understand risk.
Monitor
Monitor and track key metrics through a powerful dashboard.
Audit
Audit and act on continuous insights for reliable trust.
Our Practice Areas
We audit a wide range of AI systems for bias and fairness, data provenance, misinformation, hallucination, collusion, environmental impact and organizational readiness.
Our audit process covers a diverse range of implementations including:
- Chat: conversational models generating responses from free-form inputs
- RAG (Retrieval-Augmented Generation): integrating text generation with external information retireval for context aware responses
- Solution-based: custom models engineered to address specific challenges such as ranking, recommendations, and classification.
To ensure rigorous auditing, we apply key metrics tailored to Gen AI:
- Ethics & Safety: Diversity, Non-Discrimination, Harmful Content
- Red-Teaming: Security, Privacy and Technical Vulnerabilities
- Misuse and Technical Robustness: Misinformation and Hallucinations
- Performance, Fairness & Bias: ensure consistent delivery of high-quality, reliable outputs under diverse conditions and fairness across demographic groups
The ITACA platform provides an automated approach to assess compliance with:
- Demographic Benchmarking: Aligned data respects population distribution.
Fairness: Equality, Equity, Disparate Impact, Calibrated.
Performance: Accuracy, Performance, Recall, F1
Features: Correlation, Causality, Explainability
maturity through subscription model.
The ITACA platform is equipped with a customizable dashboard, providing developers with metrics to test new features, monitor impacts, and fine-tune for unbiased performance.
Our tailored and automated solutions help organizations ensure compliance and mitigate risks for the following legal frameworks:
- AI Risk Management Framework (RMF)
- AI Liability Directive
- Colorado SB21-169 (Restrict Insurers’ Use Of External Consumer Data)
- DSA
- EU AI Act
- GDPR
- Horizon 2020
- HR 144
- IEEE
- ISO 42001, 24368, 24027
- NIST
- NYC Law 144
- OMB M-24-18
- OMB M-24-10AI
- UNESCO AI Principles
- Tailored audits per model
- Multiple models/logics
- Tailored locations
- Tailored user licenses
- ITACA API included
- Onboarding included
Full Lifecycle AI Audit
Build trust in your models in every phase of the lifecycle with pre-processing, production, and post-processing monitoring.
What our clients are saying
David Whewell
Chief Technology Officer & Head of Ethics Koa Health
"Eticas has been an invaluable partner for us on our ethics journey, with multiple audits under our belt. I am delighted to note that we continue to make great strides as an ethical organization, and this demonstrates we take Eticas’ recommendations seriously, and see the value in championing the ethical approach to digital solutions in mental healthcare. We constantly hear from our partners, our users and our investors that providing these solutions ethically is paramount."
Nadine Dammaschk
GIZ, FAIR Forward – Artificial Intelligence for All
“Eticas has been supporting us in setting up a structured approach for assessing human rights risks in our AI-related activities in sectors as diverse as agriculture, climate change adaptation and chatbot usage. During this journey, Eticas has shown great dedication in tailoring this methodology with us to the international cooperation context, based on their AI Auditing expertise. Partners and colleagues confirmed to us repeatedly the value of the structured risk assessment. We are grateful for the collaboration with Eticas and the significant impact it has had on how we operationalize AI Ethics in our project.”
Jessica Bither
Senior Expert for Migration at Robert Bosch Stiftung GmbH
“In our work with Eticas, we have greatly valued the in-depth subject expertise of the team and the high-quality research they are able to provide on complex socio-technological questions and issue areas.”