AIEthica excels in AI governance project management for companies and institutions. We efficiently orchestrate the many moving parts of successful AI from management structure, value implementation to compliance,
ML management and continous monitoring.
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 trainings, management and communication structures, business case evaluation, data evaluation, regulatory compliance to MLOps and continous model monitoring. Depending on the current situation of the company, any governance modul may be provided separately.
Many companies already have parts of an AI governance process in place, for example data governance, codes of conduct or quality management systems. A first gap analysis reveals the needed moduls 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 like business case evaluation, data needs, 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 accounability 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 civilization. Responsibility, accountability, transparency, no harm and other key ethical values in AI are extracted from older ethical belief systems like virtue ethics or utilitarian ethics. This course for staff 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 become the most important regulation for AI systems, covering high-risk applications through extensive compliance requirements. AIEthica supports companies during the compliance process with respective legal and technical advice.
| ACTIVITIES: | Evaluation / Process & project management |
| RESULTS: | Report / Certification |
The MDR (Medical Device Regulation) is the main regulatory framework in the healthcare and medtech industry. The compliance with both regulations, AIA and MRD, is still a complex task due to lacking synchronization between the two. AIEthica guides your AI team through the process.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Compliance |
AIEthica provides regulatory / guidelines research reports for over 30 country specific regulations. We evaluate your application in the context of country specific regulations regarding compliance and gap analysis.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report |
AIEthica participates and is involved in different certifcation options and programs. We propose and report on different schemes that best apply to your application.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Certification |
An algorithmic fairness assessment evaluates potential biases in data and/or algorithms. Our ML specialists support your data scientists developing data collection and modelling methodologies aimed at creating fair algorithms, and we provide advice to governments/corporates on how to regulate machine learning.
| ACTIVITIES: | Assessment / Consulting / Setup |
| RESULTS: | Report / Consulting |
Responsible data science take steps to make data AI applications depend 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 sophisticated system engineering process for moving AI applications from R&D to productization. AIEthica has developed an accompanying ethical and responsible ML process to assure ethicality during the process.
| ACTIVITIES: | Evaluation / Research / Consulting |
| RESULTS: | Report / Certification |
Continous monitoring of data and models is required by AIA as well as other regulations. AIEthica partners with different Model Risk Management (MRM) solution providers to find the best solution for your company .
| ACTIVITIES: | Evaluation / Implementation |
| RESULTS: | Training & Consulting |
To end an AI application is more than just to push the button. The decommissioning risks must be assessed and documented, for example what happens to data records, model accessibility and interfaces to other systems. AI Ethica provides a consistent roadmap to 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 are happy to help.