Banks within the EU/EEA have reported nearly EUR 5 tn in exposures to counterparties located outside the EEA, which is 23.4% of their total exposures. Over the past year, these exposures increased significantly by more than EUR 350 bn (7.9%). A substantial portion of this growth was driven by increased exposures to counterparties based in the United States (EUR 1.3 tn in June 2024, an 11.8% YoY increase). The United States remains the largest non-EEA counterparty for the EU banking sector, followed by the United Kingdom (EUR 900 bn). Exposures to emerging markets also rose and equalled those of the UK.

Geopolitical tensions have intensified globally over the past few years amid deteriorating diplomatic relations between the United States and China, Russia’s invasion of Ukraine and war in the Middle East, and further conflicts around the globe. Although exposures to these regions are not negligible, they only account for a fraction of total exposures of the EU/EEA banking sector. Banks have almost EUR 225 bn in exposures to Middle Eastern countries, with Turkey and the UAE alone representing nearly EUR 100 bn and over EUR 45 bn, respectively. Exposures to nations directly affected by the conflict in the Middle East are limited to under EUR 10 bn, mostly towards Israeli counterparties (EUR 7 bn). Chinese counterparties account for about EUR 80 bn, while Taiwanese exposures are around EUR 18 bn. Exposures to entities in Russia, Ukraine, and Belarus are below EUR 45 bn.

Threat from terrorism is also widely elevated level, and there are increasing risks of rising trade barriers and the introduction of new tariffs. Geopolitical risks can adversely affect both the financial markets and the real economy, as well as the linkages between the two, representing a threat to financial stability. On the financial side, an increase in geopolitical risks can lead to restrictions on capital flows and payments or increased investors’ risk aversion, thereby affecting cross-border capital allocation and asset prices. This can result in heightened volatility in financial markets. On the real economy side, an increase in geopolitical risks can lead to restrictions on the import/export of goods and services or disruptions to supply chains, thereby affecting international trade and economic growth, and generating inflationary pressures. The interconnections between the two channels can potentially give rise to a detrimental feedback loop between the real economy and the financial channel, thus amplifying the overall impact of geopolitical risk on financial stability.

Geopolitical risks may also amplify existing vulnerabilities leading to situations of severe distress. These risks may have a direct and indirect impact on the EU/EEA banking sector, as counterparties of EU/EEA banks are spread across over 220 countries and have exposures to sectors potentially vulnerable to geopolitical risks. The direct impact could result from heightened credit risk associated with exposures to counterparties located in countries experiencing increased geopolitical risk. Additionally, there may be indirect effects arising from supply chain disruptions caused by geopolitical tensions and reduced demand for European products due to imposed tariffs.

The latter can be a result of political developments regarding trade policies, including tariffs, trade agreements and import/export regulations. Protectionist policies might affect international trade flows, corporate earnings and overall market sentiment. In broad, political uncertainty can create a risk-averse environment, leading to decreased investment and market activity, potentially prompting investors to seek ‘safer assets’, leading to capital flight and reduced market liquidity. Prolonged political instability can hamper economic growth, affecting not only corporate profits, but market performance too. Political uncertainty can manifest itself through governmental instability, for instance where frequent elections resulting from hung parliaments might lead to significant policy changes, or rising threat from so-called populist parties and / or politicians. These shifts can impact market expectations and investor behaviour.

These factors could negatively influence the performance of the banking sector, affecting lending volumes and practices, risk management and capital requirements. They can also increase operational costs for banks, potentially reducing profitability. Market, liquidity, and operational risks can also be impacted, for instance, during financial market turmoil, increasing margin call demands, and interconnectedness between banks and sovereigns. Potential impacts may also involve employers, offices, branches, and systems being affected by an escalating conflict or terrorist attack. Other associated risks encompass increasing ICT and cyber threats, which may involve potential complications with third-party providers, for example. The emergence of geopolitical risks can lead banks to face increasing legal and AML risks, alongside dealing with fresh sanctions and related measures that increase operational risks. These challenges also extend to correspondent banking and impact internal frameworks and processes such as prudential and accounting models, as well as valuation procedures. Banks must be adequately prepared to manage these risks.

EU/EEA banks show notable exposure to some vulnerable geopolitical risks sectors, but only limited direct exposures to vulnerable countries

The following analysis primarily addresses the aspect of credit risk, evaluating vulnerability via direct as well as – to the degree possible – certain indirect effects. An essential phase of this assessment involved identifying countries and the business sectors potentially most susceptible to geopolitical risk.

The identification of the most vulnerable countries was based on country risk scores. In concrete, S&P Capital IQ’s country risk scores were used. The score considers political, economic, legal, tax, operational and security risks that arise from doing business with or in a specific country. The scale ranges from '1' (very low risk) to '10' (extreme risk). For the purpose of the analysis, all countries classified at least as ’high risk’, i.e. those with an overall country risk score of 2.4 or above, were considered. This resulted in the selection of 82 countries, of which 10 were identified as severe-risk countries with an overall risk indicator of at least 4.4.

EU/EEA banks’ direct exposures to these geopolitically high-risk countries exceeded EUR 500 bn as of June 2024, representing around 2.5% of the total exposures of EU/EEA banks. A significant concentration of these exposures was found in Spanish banks through their subsidiary operations in Mexico and Turkey (around EUR 220 bn and EUR 57 bn), though notable exposures to higher geopolitical risk countries are present throughout several banks in Europe, albeit with a great degree of heterogeneity. In most EU/EEA countries, banks reported direct exposure to countries vulnerable to geopolitical risks of less than 4% of their total exposures. While countries such as Spain, Hungary and Croatia have slightly more such significant exposures, it was nevertheless less than 11% of their total exposures (Figure 3).

The identification of the most vulnerable business sectors was based on the simple correlation between the equity performance of relevant sectors, as indicated by selected Euro Stoxx sectoral equity indices, and the GPR Daily Index, considering a time horizon of 3 years, from mid-September 2021 to mid-September 2024. For the analysis, the five sectors that exhibited the highest negative correlation with the GPR Index were selected and assumed to be those facing particularly elevated vulnerabilities relating to rising geopolitical risks. The findings of the analysis showed that businesses related to accommodation and food service activities, transport and storage, information and communication, trade, and manufacturing were the most negatively correlated. Businesses concerning human health services and social work activities as well as construction showed an intermediate correlation.

Banks depend on a stable economy to function effectively. High volatility, caused by economic instability, political uncertainties, or monetary policy changes, leads to greater risks in lending and investments. This unpredictability can increase loan defaults and reduce asset values, harming banks’ balance sheets.

Banks maintain a diverse portfolio of assets such as loans, mortgages, and various securities. The latter are highly responsive to market dynamics. In times of market turbulence, their values can experience marked volatility, potentially resulting in substantial losses. An abrupt hike in interest rates can diminish the value of bonds and other fixed-income instruments in banks’ holdings, thereby impacting their profitability. Increased market volatility can also hinder banks’ liquidity positions and increases the liquidity risk (see Chapter 3.3).

Investor confidence plays a pivotal role in the banking industry. Turbulent market conditions can undermine this confidence, potentially triggering a sell-off in banking equities. High levels of perceived uncertainty and risk prompt investors to shift their funds towards safer assets such as government bonds or gold. This flight to safety often leads to a drop in bank stock prices, complicating efforts for banks to raise capital via equity markets.

The VIX measures the implied volatility and indicates the markets’ expectations of volatility over the next 30 days. This index is famous for reflecting investors sentiment and market uncertainty. Usually there is an inverse correlation between the index and the stock market movements, i.e. when the VIX surges the stock prices drop. The index has been mostly below 15 since the beginning of the year with the exception of a short-lived spike in March. However, since June, market volatility has increased significantly peaking at 38.6 following the ‘Carry Trade Event’. After this event, volatility levels remained elevated in comparison to the rest of 2024. Multiple factors may have contributed to this market volatility. The initial spark might have been Japan’s central bank deciding to raise its policy rate, triggering the carry trade unwind, but market participants were also concerned about various economic data from major economies that fell below expectations. This included slow growth in China and concerns over a possible recession in the United States. Moreover, central banks, particularly the Federal Reserve, hinted at potential interest rate hikes to combat inflation, raising concerns among those worried about the impacts on economic growth. Additionally, geopolitical tensions and a series of disappointing earnings reports from corporations further exacerbated market uncertainty and added to the negative sentiment. (Figure 6).

Source: Bloomberg

These market volatility spikes also impacted banks’ stock prices. In fact, European bank shares saw more pronounced price corrections, significantly underperforming the General European Index during the summer market downturns. While the June corrections in banking equities, was more evident for French banks, due to country’s political uncertainty, the price corrections were more widespread during the August market turmoil (the bank index suffered losses of 8.3% over 2 trading days, whereas the general Eurostoxx index declined by only 3%). This was likely because the recent market turmoil was closely tied to developments in interest rates  (Figure 7 and Figure 8).

Sources: Refinitiv, Bloomberg

(*) Eurostoxx (SXXE) Index and Euro Stoxx Bank (SX7E) Index which are part of the Eurostoxx Index, to secure comparable results. Eurostoxx Banks (SX7E) Index is a capitalisation-weighted index which includes listed banks in countries that are participating in the EMU.   

Figure 8: Selected European Banks Weekly changes of stock prices (%) *

Source: Refinitiv, S&P Capital IQ

(*) The banks’ sample covers different European jurisdictions but with a heavier weight towards French banks, following the political events in France. As price reference closing price of the last trade (TRDPRC_1) was used for Refinitiv data.

While volatility can create trading opportunities, it generally presents considerable difficulties for banking stocks. Fluctuating markets can reduce asset values, erode investor confidence and adversely affect profitability. Consequently, stability and predictability are crucial in the banking sector, making volatility unfavourable. Considering the episodes of heightened market volatility in 2024 and the ongoing macroeconomic uncertainty, increased caution is advised. Potential escalations in market volatility may arise from various sources, such as geopolitical tensions, political developments, and macroeconomic factors, including fiscal policy decisions and subsequent central bank monetary policies.

Banks’ exposures to NBFIs could be a source of vulnerability in periods of turmoil, as banks are often closely intertwined with different types of NBFIs. The EBA published in early summer 2024, as part of its RAR, a detailed assessment of the risks to the banking sector stemming from these interlinkages. As of June 2024, EU/EEA banks’ exposures to NBFIs amounted to 9.8% of consolidated bank assets. Large banks are generally more connected to the non-bank sector, with exposures amounting to 10.4% of total consolidated assets, followed by medium-sized banks (5.0%) and small banks (5.3%).

Exposures towards NBFIs are highly concentrated in a few countries, as 80% of the total exposures of EU/EEA banks, are reported by only five countries, (i.e. France, Germany, Italy, Spain, and the Netherlands). Of these, with the exception of Spain, all countries reported relative total exposures towards NBFI counterparties as a percentage of total consolidated assets above the EU/EEA average. In most countries, the main links to banks on the asset side are in common non-trading loans, followed by debt securities and reverse repos (Figure 10).

Source: EBA Supervisory Reporting data

The considerable increase in NBFI activity over the last 10 years is partly due to banks refining their business strategies to adapt to regulatory changes, such as CRR3/CRD6. Therefore, in some cases, these providers of ‘private credit’ could have become alternatives for bank lending. Results of a recent EBA survey with competent authorities performed as part of the regular work to monitor banking sector risks, suggests that NBFIs (particularly other financial intermediaries) in the EU/EEA largely tend to cover more niche markets that may have limited opportunities to access traditional bank financing. This includes, but is not limited to, low credit score consumer loans, specialised consumer credits (e.g. payday loans), leasing, factoring, real estate, or microfinance including SMEs.

NBFIs seem to be playing a particularly important role in the CRE market, especially as part of development and investment strategies (i.e. strategies to invest in illiquid assets). While the European CRE debt market is still dominated by traditional banks, the sharp rise in interest rates combined with lower valuations and the tightening of regulation have led to a significant reduction in banks’ overall appetite for financing in this market. In addition, the CRE market is facing secular trends, such as Europe’s focus on sustainable urban development and renewable energy, which often require specific financing requirements that are less suited to traditional bank loans and may be more suited to providers of alternative financing with more tailored arrangements, creating market opportunities for NBFIs.

NBFIs are also important for corporate lending, especially for SME financing amid rising interest rates and tightened SME lending standards by banks. Despite the continuing application of the SME supporting factor, banks face relatively strict capital requirements that limit their SME lending. Therefore, SMEs – which are responsible for a large share of employment in the EU – often face financing challenges because their access to external financing is limited for various reasons, including insufficient credit and rating history, opaque corporate structures and higher risk profiles. As far as NBFI lending is originated at fair prices and otherwise adequate standards, NBFIs which offer customised financing solutions tailored to the specific stage of SME financing could be an important complement to traditional bank financing in closing the financing constraints experienced by SMEs.

Consumer lending can be considered a more niche market in many jurisdictions, and NFBIs are in particular complementary to banks in providing services to segments of the population that have limited access to traditional bank lending for reasons such as irregular income or insufficient credit history.

Nevertheless, the results of the survey also indicate that, depending on the country, activities of NBFIs may largely overlap with those of the banking sector. Hence, banks and NBFIs are often seen as complementary to each other.

While improved access to credit as such undoubtedly improves welfare, and new types of lenders may cover parts of the market that are no longer attractive to banks, concerns have been raised. These include that lending standards may not always be as prudent as those applied by regulated financial institutions. The EBA survey also looks at how credit standards and asset quality at NBFIs compare to the banking sector. Responses suggest that NBFIs often charge higher fees and interest rates on credit products compared to banks, which could ultimately affect investment and economic growth. Main reasons for this might be NBFIs’ increased preparedness to engage in higher-risk lending, or looser and more flexible underwriting standards, with products often not meeting the standards required by regulated financial entities. As a result, this may lead to greater variability in asset quality, particularly among non-bank lenders that are not owned by banks. The potentially reduced capacity of less-regulated lenders to absorb credit losses and/or their unwillingness or inability to remain in the market during economic downturns could pose risks of a credit crunch for affected borrowers with limited access to other sources of financing. Even if in the EU the volume of NBFI lending remains moderate and is as such unlikely to be of direct systemic relevance, hidden risks may have been created which need to be carefully identified.

The first system wide climate exercise shows that transition risk losses alone unlikely to threaten EU financial stability.

As part of its 2021 Strategy for financing the shift to a sustainable economy, the European Commission requested that the ESAs and the ECB conduct a one-off ‘Fit-for-55’ climate risk scenario analysis, targeting banks, investment funds, occupational pension funds, and insurers. The objective is to assess the resilience of the EU financial sector to climate and macroeconomic financial shocks, while the Fit-for-55 package, which aims to reduce EU emissions by at least 55% by 2030, is smoothly implemented in the EU.

The ESRB, in close cooperation with the ECB, developed three scenarios, one baseline and two adverse scenarios. One adverse scenario focuses on climate-change related risks that already materialise in the near term, in the form of asset price corrections triggered by a sudden reassessment of transition risks, so called ‘run on brown’. A second scenario combines such climate-change related risks with other stress factors, as far as possible consistent with scenarios of the EU-wide stress-testing exercises. In all three scenarios the Fit-for-55 package is assumed to be successfully implemented by 2030 as planned.

The ESAs and the ECB employed top-down models to measure the impact of the scenarios on the respective sectors (first-round effects) and to assess the potential for contagion and amplification effects across the financial system (second-round effects).

The results of the exercise show that estimated losses in the near term, stemming from a potential ‘run-on-brown’ scenario have a limited impact on the financial system. This means that perceived changes in climate risks alone, as reflected in the scenarios, are unlikely to trigger financial instability. The overall system-wide losses, which include both first and second-round effects, represent 5.3% of total exposures in the baseline scenario and rise to 8.7% under the first adverse scenario. Losses specific to the banking sector account for 5.8% and 6.8% in these respective scenarios.

The interaction of adverse macroeconomic developments with climate risk factors could substantially increase financial institutions’ losses, thereby leading to disruptions in financing the ongoing transition. This is assessed in the second adverse scenario where the “run-on-brown” is coupled with adverse macroeconomic conditions. In this scenario, the losses across the entire system and also including cross-sectoral amplification effects, can reach up to 20.7% of total exposures.

Amplification effects vary widely across sectors. In the simulation, investment funds experience more severe liquidity pressures due to redemptions, leading to fire sales of assets. This mechanism affects the funds’ value and further impacts the price of securities held by financial institutions in the three sectors. Insurance corporations are mainly affected by the channel of fund share revaluations while banks have a relatively lower exposure to funds, which explains the more contained total impact (11%) compared to other sectors (Figure 16).

Source: EU-wide cross-sectoral assessment of climate-related financial risks

This exercise represents a significant step forward in the realm of climate stress testing, mainly in complexity and the incorporation of interconnected elements. Nevertheless, these estimates rely on several crucial assumptions, particularly concerning the second-round effects. Modelling uncertainty could also affect results, as the scenario construction itself involves highly detailed macroeconomic modelling. Differences in data coverage and dependence on various data sources increase the overall uncertainty of the findings. Despite inherent limitations, the exercise endeavours to maintain consistency across sectors, both in scope and methodology.

The preparedness and adaptability of institutions in managing climate-related risks can significantly contribute to drive the economy towards the net-zero target by 2050 and avoid the impacts of extreme weather events, sea-level rise, and natural disasters. To drive Europe’s green transition and ensure long-term sustainability and resilience in an unpredictable global landscape, a thorough risk management strategy that includes both transition and physical risks, along with the efficient and strategic allocation of financial resources, is crucial.

Resolution authorities submit MREL decisions to the EBA annually at end of May for decisions in force as of 1 May. The reporting covers decisions for 551 entities for which resolution authorities set MREL above own fund requirement, of which 357 are external MREL decisions and 194 internal MREL decisions. The below analysis focuses here on external MREL, considering a subsample of 339 banks (some were excluded on data quality grounds).

Bail-in is the preferred resolution tool in terms of RWAs while transfer is preferred in number of banks. The overview of the decisions received – and the resolution tools - shows that banks with bail-in as preferred tool represent about 93% of sample’s RWA. However, in terms of numbers of banks the data shows that banks with transfer as preferred tool represent half of the population of resolution banks. This reflects the fact that the transfer tool is preferred for smaller banks (Figure 33 and Figure 34).

Optionality continues to be limited for banks with bail-in as a preferred tool. BRRD requires resolution authorities to prepare variant strategies in resolution plans to address different possible scenarios, in practice this can mean considering alternative resolution strategy and tools. Looking at banks with bail-in as preferred tool, 62.8% of the banks (75% in terms of RWAs) do not have a variant strategy, while 31.8% (22.9 % in terms of RWAs) have sale of business as a variant and 5.4% (2.2% in terms of RWAs) have bridge-bank as a variant.

Over the past 6 years, US banks have generally been more profitable than those in the EU. Only in 2023 EU/EEA banks showed higher profitability, amid a decline of US banks’ RoE. The latter seems to be driven by a temporary contraction of fee and other income as well as a spike in expenses, which is mainly due to HNWI contributions amid the Silicon Valley Bank induced events last year. However, US banks’ profitability slightly surpassed EU/EEA banks’ again in 2024. Likewise, US banks have consistently reported higher return on assets (RoA). This box explores how various elements of banks' RoE might account for the consistent advantage of US banks.

One parameter in the analysis is equity, as a starting point, as the RoE’s denominator. Using capital as a proxy for comparing this parameter, data shows that EU/EEA banks’ leverage ratio has been constantly below that of their US peers, which implies that US banks’ have higher capital levels relative to their assets. This suggests that US banks need higher earnings to achieve comparable or superior RoE compared to EU/EEA banks. (Figure 62).

Looking at the two of the most important income components of the numerator of the RoE, both NII as well as fee and other income, measured as a share of equity, tend to be higher for US banks than EU banks. US banks’ NIM is consistently higher, influenced by the pricing of assets and liabilities, which depends for instance on their product mix and pricing convention (variable or fixed rate loans, high yielding vs low yielding business, etc.), deposit betas, client structure, competition in specific segments and sectors, but also the interest rate levels as set by the central banks (Figure 63).

The cost efficiency ratios paint a more mixed picture though. The cost-to-income ratio of EU/EEA banks declined notably in recent years, in contrast with the US banking sector, where it has been on a slightly upward trend. As a result, US banks’ CIR has been above the EU one since 2021. At the same time the picture of cost of risk was rather mixed and significantly more volatile, especially for US banks. Since the pandemic, cost of risk in US banks was above their EU peers, while over the last 6 years the average cost of risk for EU banks was marginally higher than US banks (54 bps vs 53 bps respectively) (Figure 64).

Source: Federal Reserve Bank of New York and EBA supervisory reporting data, EBA calculations

Source: Federal Reserve Bank of New York and EBA supervisory reporting data, EBA calculations

(*) For the US banks’ cost of risk the annualised loan loss provisions as percentage of total loans from the data the Federal Reserve Bank of New York’s US banking sector data are used.

The profitability comparison of the two banking sectors demonstrates that US banks despite their higher cost base, supported by higher NII, NIM and also fee and other income, manage to consistently fare better than EU/EEA banks. There are presumably many reasons for US banks’ higher revenues, including diversification of income, asset mix, including the use of securitisations to move certain exposures from the balance sheets, asset quality (legacy loans), but also the macroeconomic environment, including interest rates and economic growth. The market structure might similarly contribute to the differences in revenues and overall profitability. Whereas US banks benefit from a fully integrated common market at home, their EU peers do not benefit from a fully integrated banking and capital markets union, and are challenged with market fragmentation. Further aspects include for instance the political environment, prudential, liquidity and other regulatory aspects as well as technological and innovation headroom due to investments made in the previous decade by US banks.

Even though valuations are driven by many parameters and are particularly driven by the expectations related to an investment, one might still argue that different profitability levels are at least partially also reflected in respective banks’ valuations. EU banks’ valuations have been below those of their peers for years. PtB multiples provide a measure for banks’ valuations. EU banks’ PtB multiple stands at around 0.8 whereas for US banks’ it reached around 1.3 in September 2024. The trend in bank valuations over time shows that EU banks had some periods in which they could narrow the gap to their US peers. This was for instance the case at the beginning of the pandemic in 2020, when US banks’ RoE fell to nearly similar levels as EU banks’ RoE. US banks’ valuations declined significantly more than for EU banks at that time. In 2023 EU banks’ profitability outperformed that of their US peers not least due to different perceptions of the rate expectations, but also in the aftermath of the SVB induced crisis events. That time, US banks’ valuations declined amid their contraction in profitability, and the gap between EU and US banks’ valuations narrowed again (Figure 65).

Source: Bloomberg

(*) The Bloomberg query INDX_PX_BOOK was used for this analysis.

Additional, and maybe even more relevant, parameters defining banks’ valuations, include expectations regarding the general economic and monetary policy and interest rate environment, and how this affects banks’ profitability going forward. Investors’ investment strategies, such as  focusing on dividend investing or  value investing (i.e. investing in stocks that seem to offer an increase in their value) can  also affect valuations of certain sectors. Another factor, could for instance be windfall taxes for banks that might have affected the valuations of certain banks in recent quarters.

Transition to a more sustainable economy has resulted in increased demand and supply of sustainable products in recent years. One of the side effects of this change is increased risk of greenwashing, which is now receiving more attention with the potential to impact the transition by reducing investor confidence and necessary investments that support meeting the objectives established by the European Green Deal. It can also generate reputational and financial risks for the institutions, including through litigation processes, and can affect the overall credibility of sustainable finance markets. However, the negative effects of greenwashing are not only limited to operational and reputational risks but can also have an impact on liquidity and funding risks and affect the whole business model, which is why the EBA has taken a deeper look into it in the last few years and continues to monitor it.

The EBA’s Final Report on greenwashing published in June 2024 confirmed the observed trend – a clear increase in the total number of potential alleged cases of greenwashing across all sectors. The overall number of alleged greenwashing cases was 7.3 times higher in 2023 compared to 2012. In 2023, the total number of alleged cases increased by 21.1% globally and 26.1% in the EU compared to 2022. Looking at the EU financial sector, the increase observed over the last decade in terms of the absolute number of alleged greenwashing cases are far lower than in EU non-financial sectors. While in 2012 the EU financial sector accounted for 28% of all alleged greenwashing cases involving an EU company, of which 19% were at EU banks, the shares decreased to 21% and 8% respectively in 2023 (Figure 72). 

The EBA has a statutory duty to monitor and assess market developments, including financial innovation, and to achieve a coordinated approach to the regulatory and supervisory treatment of new or innovative financial activities. In accordance with this mandate and the EBA’s priorities on innovative applications for 2024/2025, that include AI/machine learning, including GPAI, the EBA is deepening its monitoring and analysis of the uses of GPAI by EU banks. Since September 2023, the EBA has collected data on the testing and use of GPAI by EU banks via the EBA’s RAQ. The results have provided the EBA with detailed insights into the levels of adoption and uses of GPAI in the EU banking sector.

Additionally, in April 2024 the EBA organised a Workshop on GPAI in the banking sector, which included the participation of a range of EU stakeholders representing banks, consumer organisations and technology providers, in addition to relevant EU and national competent authorities. The workshop aimed to promote a common understanding of the uses, risks and opportunities associated with GPAI in the banking sector. During the workshop the EBA found that the use cases of GPAI are rather limited, with activity mainly focused on testing around a small number of use cases. The EBA found that the banking sector has limited tools to fully mitigate consumer protection concerns associated with GPAI, which justifies the cautious and gradual approach to the adoption of this technology.

Based on engagement with competent authorities, market stakeholders and consumer organisations, the EBA has identified that the following consumer protection issues could arise if banks adopt GPAI in consumer-facing use cases:

Accountability. The complexity and limited transparency associated with GPAI could raise challenges in terms of ensuring that providers of GPAI systems and models to banks are accountable for inaccurate or inappropriate outcomes, such as inaccurate, misleading or false information or advice to bank customers. However, as banks are accountable for inaccurate, misleading, or false information provided to consumers according to general financial protection legislation, consumers may seek compensation for harms through banks. 

Bias, discrimination and financial exclusion. Training data used by GPAI model developers can, without mitigation, exacerbate discrimination and bias against minority or misrepresented groups, leading to risks of financial exclusion. Due to the randomness associated with GPAI outputs, banks deploying GPAI in customer-facing applications may face limited capabilities for understanding and explaining the logic behind biased outputs. Additionally, for instance, consumer support for speakers of minority languages may be impacted if such support gradually shifts from human to GPAI-powered virtual assistants, since GPAI models are mainly trained on the largest languages.

Transparency. The opacity of GPAI applications implies that informing consumers about the use of GPAI in consumer-facing interfaces may not be sufficient for raising their awareness. Therefore, banks might need to explore the use of other means (e.g. application-specific disclaimers) to provide more complete information to consumers.

Data security. The improved technical abilities of attackers (including new types of cyber attacks, prompt-related attacks, misinformation campaigns, data poisoning attacks, and data exfiltration techniques) can raise data security concerns for consumers. Use of GPAI can also contribute to data security risks for consumers via information leakage or inappropriate or non-factual responses. However, thanks to GPAI’s potential benefits in areas such as programming script analysis or malware detection and investigation, existing research suggests that there is no substantial evidence yet suggesting that GPAI can automate sophisticated cybersecurity tasks which could tip the balance between cyberattackers and defenders in favour of attackers.

Other concerns. As GPAI can potentially change the threat landscape for banking consumers by boosting cyberattackers' technical skills, aiding social engineering or phishing attacks, and speeding up manipulation techniques through consumer personification and hyper-realistic scams, GPAI could introduce challenges in consumer experiences.