What CFA Institute's new AI framework means for Canadian portfolio managers and institutional investors.
Summary prepared by CFA Montréal's Fintech Committee, based on the report Artificial Intelligence and the Future of Finance: A Framework for Structural Change, by Mona Naqvi (CFA Institute, July 2026).
Artificial intelligence is redefining the very terms on which investment professionals practise their craft, far beyond being just another tool to evaluate. That's the argument Mona Naqvi makes in Artificial Intelligence and the Future of Finance: A Framework for Structural Change, published in July 2026 by the CFA Institute Research and Policy Center, the first paper of its kind to treat AI as a structural force rather than simply a question of adoption. The Fintech Committee is opening the season with this topic for a simple reason: Canadian regulators have just made the issue very concrete.
Mass literacy, meaning the spread of reading and writing throughout society, reshaped the very organization of society by making public education and mass democracy possible, well beyond any simple productivity gain for businesses. The same will hold true for artificial intelligence and capital markets.
Capital markets rest on a simple premise: informational scarcity, in the right place at the right time, provides a competitive edge. Artificial intelligence upends that logic. As analytical capacity becomes widespread and gets built into market infrastructure, intelligence turns abundant, and it's judgment about how to govern it, deploy it and rein it in that becomes the scarce resource. Whether AI adoption will continue is no longer really the question; what remains to be seen is how capital markets will reorganize around this new analytical abundance.
Five foundational principles of modern finance are under strain, now that the conditions of scarcity that underpinned them have disappeared.
The report outlines four possible configurations, ranging from a straightforward productivity gain to markets fully mediated by models, and notes they could coexist from one firm to the next. A common tension runs through all of them: once analytical tools become accessible to everyone, differentiation shifts from producing information toward system design, data quality and consistency of execution. For an active Canadian manager, the implication is direct. The report describes a strategic fork with three paths: compete on scale, specialize where human judgment still stands out, or differentiate through governance. Few firms here will out-scale the global platforms. The other two paths remain open, but the choice needs to be made now.
The report's most original concept is cognitive convergence: the gradual alignment of model architectures, training data and decision-making frameworks across institutions. The result is correlated analytical outputs that squeeze out the interpretive diversity that market resilience depends on. Traditional prudential frameworks, built around the solvency of individual institutions, aren't designed to catch this. The stakes are especially high in Canada, where most analytical infrastructure is leased from a very small number of foreign vendors.
Three priorities stand out: traceability of decisions within hybrid systems, validation frameworks suited to evolving architectures, and oversight of shared analytical infrastructure. The risk of overregulation is real. The risk of under-governing a transition already underway is greater. Here, the framework is already taking shape.Canadian Securities Administrators (CSA) Staff Notice 11-348 offers guidance on how Canadian securities law applies to registrants' use of AI systems, with an emphasis on governance, oversight and explainability. OSFI's Guideline E-23 on model risk management takes effect May 1, 2027 for federally regulated financial institutions, and theAutorité des marchés financiers (AMF) has become the first provincial regulator to publish its expectations.
This report isn't announcing the end of active management, but a shift in what it's judged on. Judgment is migrating from analysis to overseeing systems, and the real question becomes: what will a manager count on to stand out once analysis is commoditized? The risk that's least well covered is systemic, not firm-specific: model convergence and vendor concentration. Path dependency is real: governance decisions made early lock in, or lock out, options for years to come. Choosing not to decide is still a decision. Governance credibility will become a criterion for winning and keeping mandates, not just a compliance requirement. That's the conversation the Fintech Committee is opening this season.
References
CSA Staff Notice and Consultation 11-348, Applicability of Canadian Securities Laws and the use of Artificial Intelligence Systems in Capital Markets, December 5, 2024.
OSFI, Guideline E-23, Model Risk Management (2027), final version published September 11, 2025, effective May 1, 2027.
Autorité des marchés financiers (AMF), Guideline for the Use of Artificial Intelligence, effective April 24, 2026.
Mona Naqvi, Artificial Intelligence and the Future of Finance: A Framework for Structural Change, CFA Institute Research and Policy Center, July 2026.
Latest news
