The AIEthica AI Governance Framework is a guided, flexible and modular step-by-step approach to responsible AI. Modules range from general impact assessments, policy training, management and communication structures, business case evaluation, data evaluation, regulatory compliance to MLOps and continuous model monitoring. Each governance module can be delivered separately, depending on the company's current situation.
Many companies already have parts of an AI governance process in place, such as data governance, codes of conduct or quality management systems. A first gap analysis reveals the necessary modules for the further establishment of an effective AI governance process.
| ACTIVITIES: | Evaluation / Analysis |
| RESULTS: | Report / Consulting |
In the design phase of the AI system, the analysis covers topics such as business case evaluation, data requirements, models and features, problem/system fit, cost/benefit analysis, risk analysis, metrics as well as stakeholder analysis.
| ACTIVITIES: | Evaluation / Analysis |
| RESULTS: | Report / Consulting |
Governance is the establishment of accountability where all responsibilities in the AI process are clearly defined. This module maps existing management structures to AI management needs and identifies missing accountability structures in the AI process.
| ACTIVITIES: | Evaluation / Analysis |
| RESULTS: | Report / Consulting |
Training for staff and management on the ethical principles and legal policies implicated in the deployment of AI. Preferably conducted before deplyoment, but possible
| ACTIVITIES: | Course / training |
| RESULTS: | Training |
The basic ethical ideas applied to AI are almost as old as Western civilisation. Responsibility, accountability, transparency, no harm and other key ethical values in AI are drawn from ancient ethical belief systems such as virtue ethics or utilitarian ethics. This course for employees and/or managers is a deep dive into the history of ethical thinking.
| ACTIVITIES: | Course / training |
| RESULTS: | Training |
In Europe, the EU AI Act will be the main regulation for AI systems, covering high-risk applications through extensive compliance requirements. AIEthica supports companies in their compliance process by providing legal and technical advice.
| ACTIVITIES: | Evaluation / Process & project management |
| RESULTS: | Report / Certification |
The MDR (Medical Device Regulation) is the main regulatory framework for the healthcare and medtech industry. Compliance with both regulations, AIA and MDR, is still a complex task due to the lack of synchronisation between the two regulations. AIEthica guides your AI team through this process.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Compliance |
AIEthica provides regulatory / guideline research reports for over 30 country specific regulations. We can assess your application against country specific regulations for compliance and gap analysis.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report |
AIEthica participates and is involved in a number of certification options and schemes. We can recommend and report on the most appropriate scheme for your application.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Certification |
An algorithmic fairness assessment evaluates potential biases in data and/or algorithms. Our ML specialists can assist your data scientists in developing data collection and modelling methodologies to create fair algorithms, and we can advise governments/companies on how to regulate machine learning.
| ACTIVITIES: | Assessment / Consulting / Setup |
| RESULTS: | Report / Consulting |
Responsible data science takes steps to ensure that data AI applications rely on findable, accessible, interoperable and reusable (FAIR) data, while ensuring the fairness, accuracy, confidentiality and transparency (FACT) of the algorithms and tools that are created.
| ACTIVITIES: | Evaluation / Consulting |
| RESULTS: | Report |
Machine Learning Technology Readiness Levels (MLTRL) is a sophisticated systems engineering process for moving AI applications from R&D to productization. AIEthica has developed an accompanying Ethical and Responsible ML process to ensure ethicality during the process.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Certification |
Continuous monitoring of data and models is required by the AIA and other regulations. AIEthica partners with a number of Model Risk Management (MRM) solution providers to find the best solution for your firm.
| ACTIVITIES: | Evaluation / Implementation |
| RESULTS: | Training & Consulting |
Ending an AI application is more than just pressing the button. Decommissioning risks must be assessed and documented, including what happens to data records, model accessibility and interfaces to other systems. AIEthica provides a consistent roadmap for decommissioning.
| ACTIVITIES: | Evaluation / Consulting |
| RESULTS: | Consulting |
Whether you're starting an AI project, need training for your employees or want to implement model risk management, we're happy to assist you. Just contact us.