Join the Research Institute in Data Science and Artificial Intelligence (RIDSAI) at our upcoming seminar on data privacy and security in AI advancement.

The event will feature short presentations by UNB researchers followed by refreshments and a networking opportunity.

Speaker: Cathy Simpson (TechImpact CEO)

Cathy Simpson is an Atlantic Canadian IT leader who has dedicated her career to building the region's tech capacity. Most recently, she was the Vice President, People and Culture with T4G Limited, one of the largest privately held IT firms in the region with clients throughout North America.

Cathy is a leader in the movement to use technology and innovation to drive economic growth, create thriving communities and engage citizens. She champions this evolution as Chair of the New Brunswick Innovation Foundation, an independent, not-for-profit corporation that invests in new growth-oriented companies and applied research activities.

Her current work is an extension of her role as one of the co-founders of PropelICT, Atlantic Canada’s original tech accelerator, and she continues to mentor and advise start-up entrepreneurs in the region.

In 2015, Cathy founded UP+GO, a social enterprise building and delivering leadership programs to high school girls, teaching job ready skills, strengthening their leadership mindset, and exposing them to incredible women in STEM, entrepreneurship and leadership.

She began her career at NBTel and Innovatia where she helped launch Canada’s first IPTV service and other innovative consumer products. Cathy is a graduate of F.C. Manning School of Business at Acadia University and lives with her family in Rothesay, NB.

This presentation explores the critical issue of adversarial AI and its potential to exploit AI systems, leading to significant financial and reputational risks for organizations. We advocate for a two-pronged approach to protect the integrity of AI model behaviour: rigorous penetration testing of AI models to identify vulnerabilities and targeted monitoring of model inputs to detect and mitigate potentially harmful manipulations. We examine the vulnerability of both traditional prediction-based models and newer generative AI models to adversarial attacks, emphasizing real-world examples and enterprise case studies.

Speaker: James Stewart (CTO TrojAI)

Dr. James Stewart is CTO of TrojAI, which provides cybersecurity protections for AI driven systems. Prior to this role, James held the position of SVP of Video Analytics at Patriot One Technologies (TSX:XTRA), a Canadian AI company that acquired his video analytics company, EhEye.

James brings a blend of deep learning and cybersecurity experience having held positions as Lead Data Scientist with Raytheon Cyber and Research and Development Manager with the Canadian Institute for Cybersecurity. James received his Ph.D. in computer science from the University of New Brunswick and currently serves as an international expert representing Canada for ISO/IEC SC 42 (AI).

It has become a widely accepted best practice for faculty to incorporate Generative AI tools into their courses and assessments. However, this integration raises concerns about data safety. When students use these tools, where does their data go? Is it stored or used to train future models? Can students opt out if they have privacy concerns? What are the safest ways to integrate Generative AI into classes without compromising security?

This presentation will address these questions and highlight the need to balance the benefits of Generative AI with strong data protection measure. We will examine case studies from universities that have navigated the challenges and successes of building their own Generative AI tools, highlighting the important need for clearer governance by higher education institutions in this area.

Speaker: Aya Bouchnak (UNB Instructional Designer)

Aya Bouchnak is an Instructional Designer at the Teaching and Learning Services (TLS) within the Centre for Enhanced Teaching and Learning (CETL). She holds an MSc in Instructional Design and Technology and has spent the past three years experimenting with generative artificial intelligence (AI) tools and developing resources that promote healthy AI usage in higher education, while also supporting faculty in effectively incorporating AI tools into their teaching practices.