Yapı Kredi Technology- TRUstworthy huMAN-centric artificial intelligence

Project Objective

Improving the reliability of artificial intelligence systems

It is aimed to produce safe and ethical solutions while protecting the privacy of users and the fairness of systems.

Developing human-centered artificial intelligence solutions

Transparent and explainable solutions enabling user control over the system will be designed.

Developing decision support systems with the HTIL (Human-in-the-loop) approach

Ethical decision support systems involving human participation and considering user impacts will be developed.

Developing artificial intelligence models applicable to different sectors

Generalizable models that can be used in fields such as healthcare, finance and marketing will be produced.

Developing general artificial intelligence principles and guidelines

Guidelines will be prepared by establishing fundamental principles centered on ethics, reliability and transparency.

Project Subject

Explainable and robust artificial intelligence systems that focus on user participation and security, applicable across different sectors, will be developed in the project.

Explainable and Participatory Artificial Intelligence Systems

User-centered systems will be designed with infographics, HITL principles and dynamic data collection scenarios.

Security, Privacy and Resilience Solutions

Protective approaches against adversarial attacks and robust structures developed for LLMs will be addressed.

Sectoral Applications and Policy Guidelines

Technology tests will be conducted using sample scenarios for fields such as marketing, finance and healthcare, and policy recommendations will be prepared.

What Do We Do?

Detection of In-branch Misconduct with DERIN

Unique AI solution that detects customer service fraud
Explainable and reliable system infrastructure

Adversarial Robustness and Security Solutions

Robust models tested against attacks
AI security specific to critical banking applications

Developing Ethical and Explainable AI Models

Transparent and explainable interfaces for user trust
Design compliance with Trustworthy AI principles

Introducing Regulation-Compliant AI to the Financial Sector

Explainable financial models
Compliance with global ethical standards

Secure, transparent and explainable artificial intelligence guidance for the financial sector

Practical solutions for ethical, secure and explainable artificial intelligence systems for banking and finance applications will be developed through this project. Project outputs will guide the sector in terms of artificial intelligence design, user security and compliance with regulations.

Explainable and Ethical Compliance Guidance

The artificial intelligence systems developed as part of the project will be designed to make it easier for users to understand the decision-making process. Explainability methodologies and user-friendly interfaces will be developed with a social science perspective. Ethical risks will be reduced through designs compliant with trustworthy AI principles.

Robust and Secure Artificial Intelligence Models

Artificial intelligence solutions resilient to adversarial attacks were developed, and will be tested with performance metrics. Sector-based feedback was collected through open innovation calls, and solutions will be expanded. Robust AI architectures powered by infographics and large language models will be designed.

International Collaborations and Information Transfer

Information sharing will be achieved through joint R&D processes with countries such as Canada and India. Results will be shared in accordance with the success criteria defined by the European R&D network. The project outputs will be a model for compliance with regulations and technology transfer in the sector.