Concordia partners with National Bank of Canada to advance trustworthy artificial intelligence systems

Companies now use artificial intelligence (AI) to assist with software development, where the AI can write parts of the code, test it and even fix errors. This approach not only speeds up the process but also allows human developers to focus on more complex and creative tasks.

However, integrating AI into existing processes isn't without its challenges. AI-generated code can sometimes be unpredictable or contain errors that are hard to detect. Additionally, the complexity of AI systems — which often rely on vast amounts of data and numerous software libraries — makes ensuring their reliability and security a daunting task, even more so in the financial sector.

To address these issues, Concordia University, in partnership with National Bank of Canada (“National Bank”), has received $1.5 million over five years to improve the development of trustworthy AI-based systems. This initiative, led by Emad Shihab, a professor at the Gina Cody School of Engineering and Computer Science, is co-funded by the Natural Sciences and Engineering Research Council of Canada (NSERC).

“This collaboration aims to create actionable knowledge and advanced technologies essential for the future of AI,” Shihab says. “The research will not only benefit National Bank but also provide valuable insights and solutions for the broader AI community.”

National Bank, one of Canada's six systemic banks, invests in AI to improve the client and employee experience and is an active player in the Quebec, Canadian and international ecosystem surrounding responsible AI. The bank's active participation in the project includes providing expertise and resources, ensuring that the research is grounded in real-world industry needs.

“Trustworthy AI systems are at the core of a successful bank of the future,” said Julien Crowe, senior director, artificial intelligence, at National Bank. “I am excited to tackle this complex applied research challenge in collaboration with Professor Shihab and the amazing team at Concordia.”

Automating and enhancing software development with AI

A key focus of this research is to harness the power of AI to automate and enhance the efficiency of software development processes and teams. Modern software development is a complex and time-consuming task, often requiring developers to write extensive code, test it for errors and continuously update and maintain software.

AI-based systems can significantly streamline these processes by automating routine tasks. For instance, AI can assist in generating code snippets, conducting automated testing and even identifying and fixing bugs. This not only speeds up development but also allows human developers to focus on more strategic and innovative aspects of their work.

However, AI-generated code and automated processes must be thoroughly vetted to ensure they meet quality standards and avoid introducing new issues. Moreover, the efficiency gains from AI must be carefully measured to validate their impact on overall productivity.

As such, the project will develop innovative techniques to seamlessly integrate AI into various stages of the software development lifecycle.

Building trustworthy AI-based software

Another crucial aspect of this research is ensuring that AI-based software remains trustworthy from its initial development to its deployment and maintenance. AI systems often rely on data and code from external sources, which can introduce risks if not properly managed. For example, an update to an AI system could inadvertently introduce errors or vulnerabilities that affect its performance.

The project will develop techniques to verify the trustworthiness of AI-generated software components and ensure they remain secure and reliable throughout their lifecycle. This includes creating methods to track and manage software updates, identify potential issues and come up with tools to help developers maintain high standards of security and reliability.

Training the next generation of AI experts

The proposed program will train more than a dozen graduate students, including a postdoc, PhD students and master’s students. They will gain knowledge and experience in developing AI-based software systems, particularly in the era of advanced foundation models like large language models.

The partnership between National Bank and Concordia will additionally provide a rich training experience. Students will work on real-world problems, learning state-of-the-art techniques to address issues in AI-based software development. The program includes opportunities for internships with National Bank, giving students hands-on experience and the chance to work directly with industry professionals.

Learn more about Concordia’s Department of Computer Science and Software Engineering.