At the inaugural NextGen: AI 2024 event in London, Dr. Joe Lyske, co-founder and managing partner of Time Machine Capital, delivered a compelling keynote address titled, “What can we do with AI?”
Dr. Lyske emphasized the game-changing capabilities of artificial intelligence in the banking sector, detailing how his firm leverages AI to stress-test various scenarios and generate detailed solutions. While the integration of AI across diverse industries heralds numerous benefits, he also cautioned about the ethical dilemmas that may arise, particularly regarding the replication of human behavior and the potential to overshadow human insight.
“We are on the brink of a new era—the intelligence revolution. We are transforming data into information, and then into knowledge, all processed with wisdom to forecast outcomes, mitigate risk, and enhance certainty. The essence of intelligence lies in predicting the future. How does this perspective reshape our understanding of what AI can achieve?” he asked.
In discussing the role of AI in risk management, Dr. Lyske highlighted its potential to evaluate and manage risk, which could significantly influence compliance measures and product development. He elaborated on innovative approaches like tokenizing common behaviors and utilizing AI to monitor cryptocurrency wallets and social media for effective Know Your Customer (KYC) and Anti-Money Laundering (AML) practices. He also explored the transformative implications of Web 3.0 on human interactions, data ownership, and institutional relationships.
According to Dr. Lyske, while traditional banking relies heavily on trust, consumers are increasingly dissatisfied with conventional trust models. Instead, they are turning to technologies that empower them to operate in a trust-less environment. “This paradigm shift underpins Web 3.0 as a social, political, and economic movement,” he asserted, also noting that cryptocurrencies have addressed this shift.
He foresees a future enabled by AI, where global identities can be authenticated, facilitating seamless cross-border transactions. He articulated a vision where banks could serve as ‘blockchain oracles,’ bridging various blockchain systems by integrating real-world inputs. In this framework, banks become pivotal in verifying customer identities and legitimizing transactions, utilizing AI to ensure accuracy and security.
The event’s first panel discussion, titled “Where next with AI?”, featured expert insights from Daniel Szmukler, director of the Euro Banking Association, alongside panelists Dr. Jochen Papenbrock from NVIDIA and Jeff Tijssen of Bain & Company.
Szmukler expressed confidence in AI’s potential to transform industries, citing its ability to resonate on a personal level across generations. “AI’s tangible applications mean it touches everyone and is uniquely positioned to deliver value.”
In addressing challenges related to AI adoption, Tijssen stressed the financial industry’s need for nearly flawless accuracy. “For regulatory compliance, achieving 97% or even 99% accuracy is insufficient; we must strive for 100%,” he stated. He highlighted the critical issue of explainability, which complicates customer relations and stakeholder engagement.
Dr. Papenbrock concurred, adding that the success of AI hinges on robust technological foundations. “AI requires considerable infrastructure. At NVIDIA, we conceptualize this as an ‘AI factory’—where financial intelligence is cultivated by humans, for humans, incorporating AI wherever feasible. This ‘AI Center of Excellence’ is crucial for managing the intricacies of large models efficiently.”
He further noted that failing to establish a regulatory framework poses significant challenges for banks seeking to harness AI’s potential. The panelists agreed that cultural shifts within organizations are vital for fostering innovation and growth.
Szmukler pointed out the inherent conservatism in banking culture, which prioritizes customer trust as its most valuable asset. He warned that if AI systems operate as ‘black boxes’—lacking transparency in decision-making—this could lead to considerable reputational risks for banks in light of public scrutiny.
To support AI adoption, Dr. Papenbrock suggested the importance of having an interoperable platform capable of processing large-scale data efficiently and highlighted the need for dedicated environments to test AI models to ensure their safety and accessibility.
In conclusion, the panel underscored the necessity of explainable AI to mitigate bias and advocate for comprehensive education and training in AI adoption. Tijssen emphasized that while theoretical knowledge is vital, executive commitment is essential to drive meaningful AI integration.
Dr. Papenbrock also shared that the Deep Learning Institute provides complimentary courses, and NVIDIA is dedicated to democratizing AI accessibility through educational initiatives.