Financial System Survey highlights—2026

The Bank of Canada conducts the Financial System Survey (FSS) annually to solicit the opinions of senior experts in risk management in organizations active in the financial sector.1 These experts provide their views on the risks to, and the resilience of, the Canadian financial system as well as on new developments they are monitoring. The survey results are a useful benchmark for comparing Bank views and analytical work with outside opinions. Bank staff also use these results to identify new topics for research and analysis.

The 2026 FSS, completed by 54 respondents, took place between February 23 and March 13, 2026 (Chart 1). As a result, the survey captured the heightened volatility in several financial markets that followed the onset of the war in the Middle East on February 28.2

In addition to the set of recurring questions, this survey included a group of questions on the use and risks of artificial intelligence (AI). The questions focused on the:

• extent to which financial system participants use AI and for what purposes
• risks that using AI may pose to organizations and limitations in managing those risks
• main risks that AI use could pose to the financial system

Highlights

• Respondents believe the likelihood that a shock could impair the Canadian financial system has decreased since the 2025 survey, and their confidence in the resilience of the system remains high.
• The top risks to respondents’ organizations are similar to those from the 2025 survey. International economic and political risks remain the most significant, with cyber incidents ranking second.
• Nearly all respondents reported using AI, with most citing limited or moderate use across several business functions. The most common uses of AI among respondents are for information gathering and analysis and to support internal operations.
• Respondents generally view AI as a tool to complete existing tasks faster but not as a replacement for human judgment given the significant financial, legal and reputational risks involved.
• Going forward, respondents plan to expand their use of AI across several business functions, particularly investment management and research, and to support operational workflows and back‑office activities, financial‑crime prevention and customer service.
• Respondents identified a broad range of AI‑related risks to their own operations, including data quality and bias, cyber security and data privacy, and model risk—namely the risk of errors arising from incorrect model design, implementation or use as well as limited explainability.
• From a financial stability perspective, respondents generally viewed AI less as a standalone risk and more as a factor that could amplify existing vulnerabilities. Survey results reveal respondents share similar areas of concern about AI for the financial system and for their own operations, including model risk, cyber risk and data security. However, respondents focused on whether these issues could become more widespread across institutions or interact in ways that affect the resilience of the broader financial system. Respondents also mentioned the risks of AI adoption outpacing governance and regulation, concentration among a small number of third‑party AI and cloud providers and operational resilience, including the ability to maintain critical functions and recover from disruptions.

Risks to the financial system

Overall perceptions of risk and confidence

As in previous surveys, respondents were asked about the possible occurrence of a shock that could impair the financial system. For both the short term (less than one year) and medium term (one to three years), the perceived likelihood of such a shock occurring has decreased from that reported in last year’s survey (Chart 2).

The most common potential shocks noted by respondents related to changes in trade policies—particularly the upcoming review of the Canada‑United States‑Mexico Agreement (CUSMA)—and the risk of broader escalation of geopolitical conflicts. Respondents were also concerned about:

• the real estate sector, including mortgages being renewed at higher interest rates
• AI, including potential negative impacts of broader AI adoption on employment and the overall economy
• growth of private markets, including for private credit

Nonetheless, respondents remain confident in the resilience of the Canadian financial system (Chart 3). As in previous surveys, respondents remain confident because of the:

• financial system being well regulated, including the well-capitalized banking sector
• resilience of the Canadian financial system in previous episodes of turmoil

Some respondents continue to expect that regulators, central banks and governments would intervene if a large shock were to occur.

Most important risks

Respondents ranked the top three risks that would have the most severe impact on their organization if those risks were to occur over the next three years. They also assigned each of these risks to a broader category. Chart 4 shows the top risk categories in the order of their risk index, calculated by combining the ranking of each risk category weighted by its frequency among responses.

The top three risks in the 2026 survey are:

1. International economic and political risks. These primarily involve the risks of geopolitical conflicts, trade fragmentation—especially around the upcoming CUSMA review—and spillovers from a weaker economic outlook in the United States and globally. According to respondents, these risks could trigger a broad repricing in financial markets, higher inflation and weaker economic growth. Respondents are managing these risks by:
   • strengthening their scenario analysis and stress testing
   • maintaining sufficient capital and liquidity buffers
   • diversifying exposures and funding sources, and increasingly hedging risk exposures
2. External risks. These mainly consist of the risk of a cyber incident that affects operations, reputation or data security. Respondents are managing the risk of cyber incidents by:
   • strengthening their monitoring of cyber security risks, including evaluating risks from third‑party vendors
   • investing in cyber security initiatives
   • improving their preparedness to respond to a cyber incident
3. Asset pricing risks. These include the possibility of an abrupt repricing across financial markets, which could result in losses and liquidity strains. Respondents reported risk management approaches that were similar to those used for international economic and political risks. Several respondents noted difficulties in mitigating this risk through diversification, given recent cases of traditional safe‑haven assets, such as the US dollar, declining in value along with risky assets during periods of turmoil.

New developments

Respondents reported new developments that their organization started monitoring within the past 12 months and shared how these may affect risks to their organization. The most frequently reported new developments were those related to AI as well as geopolitical and trade tensions, both of which are discussed in other parts of this report. Other new developments included the following:

• Private markets
  • Some respondents raised concerns about illiquidity, noting that private market investors are increasingly unable to exit positions or redeem their shares from private credit funds. This illiquidity could force investors to increasingly sell public market assets to raise cash, which could present greater risks to the valuations and liquidity of public assets.
  • Other respondents raised concerns about the transparency around the holdings in private funds, noting limited visibility into the valuation practices and exposures of these funds.
• Quantum computing
  • Although some respondents pointed out that commercially viable quantum computers may still be a few years away, they said the technology is advancing rapidly enough that it can pose a threat to existing encryption standards in the future.
• The US dollar as a safe‑haven asset
  • Some respondents noted that the US dollar was no longer reliable as a safe‑haven asset during periods of turmoil and that they were looking for other types of safe‑haven assets.
  • Others noted their perception of a decline in US exceptionalism, questioning the Federal Reserve’s independence, rising US debt levels and high uncertainty around US policy decisions more broadly.
• Payment sovereignty
  • Respondents raised concerns that reliance on foreign payment infrastructures and the potential dominance of foreign‑controlled stablecoins could leave Canada vulnerable to economic coercion and erode control over the Canadian financial system.

Artificial intelligence in the Canadian financial system

This year’s survey asked respondents about their use of AI and the associated potential risks to the Canadian financial system. We asked market participants to tell us:

• whether their organization uses AI and for what purposes, as well as any planned changes to their AI use
• what risks using AI might pose for their organization, and the limitations they face in managing those risks
• how AI use could pose risks to the financial system

Current use of artificial intelligence

Nearly all respondents reported using AI, with most reporting limited or moderate use across a wide range of business functions. The most common AI applications are for information gathering and analysis and to support internal operations (Chart 5). Examples include extracting data from financial documents, summarizing news and research, automating document workflows and supporting internal analytics.

Respondents said they use both internally developed and externally sourced AI tools, including commercial platforms and chatbot subscriptions provided by major AI vendors.

Respondents generally use AI to complete existing tasks faster. They do not use it to replace human judgment or to fully automate critical decisions given the significant financial, legal and reputational consequences of relying entirely on AI.

Respondents consistently cited improvements in efficiency and productivity as benefits of AI. The time and cost savings from AI allow respondents to reallocate staff toward higher‑value activities. Other benefits mentioned were better insights and improvements in risk management. Many respondents have not quantified all of these benefits. And a few highlighted how the high costs associated with implementing AI—such as investments in data infrastructure, governance, validation and oversight—make returns on investment unclear.

Future plans for artificial intelligence use

Over the next two years, market participants intend to expand their use of AI across several business lines (Chart 6). Respondents most often mentioned plans to increasingly use AI for researching and managing investments, supporting operational workflows and back‑office activities, preventing financial crime, improving and enhancing efficiencies for customer service, and generally improving employee‑level productivity.

Responses also varied depending on the type of respondent:

• Investment fund managers and pension funds frequently reported plans to use AI to aid in market research, leverage big data to inform investment risk models and enhance monitoring of exposures and risks.
• Banks, broker‑dealers and credit unions intend to implement AI broadly across all business functions, including operational process improvements, financial crime prevention, risk management and stress testing.

Respondents acknowledged that they may face hurdles in expanding their use of AI. Among respondents:

• 58% cited difficulty integrating AI into existing infrastructure and workflows
• 56% reported talent‑related constraints, pointing either to insufficient AI literacy skills among existing staff or difficulties hiring and retaining employees with AI‑specific expertise
• 33% reported data security and privacy concerns
• 31% cited high costs of implementation and use

Artificial intelligence and the risks for organizations

Respondents identified a broad range of AI‑related risks to their own operations, but three risk categories dominate (Chart 7):

• data quality and bias
• cyber security and data privacy
• model risk or lack of explainability

Respondents also highlighted the lack of robust backup plans if AI‑enabled systems fail.

Among respondents who noted challenges to risk management, most cited limitations in governance maturity, skills shortages and AI literacy gaps. Respondents also spoke about AI technology evolving more rapidly than best practices, industry standards and risk management practices (Chart 8). Overall, responses suggest that the main challenge is developing the institutional capacity to effectively manage the risks from AI as its use expands.

Risks to financial system stability from artificial intelligence

Respondents generally view AI less as a standalone source of financial stability risk and more as a risk amplifier that could intensify existing vulnerabilities (Chart 9). Respondents frequently noted risks related to:

• Model risk and opacity. These include concerns about overreliance on complex automated models, embedded bias, difficulty tracing decisions back through opaque model logic and the risk of shared mistakes if financial institutions use similar models.
• Cyber security and fraud. AI could accelerate the volume and sophistication of cyber attacks on financial institutions and infrastructure.

In addition, respondents pointed to risks arising if AI adoption outpaces governance and regulation; concerns about concentration in third‑party AI and cloud providers; and operational resilience in the event of outages, failures or disruptions at financial market participants or with critical financial infrastructure.

Endnotes