Blog

1. What’s a conversation we’re not having enough when it comes to AI and the future of work? 

There is no shortage of headlines and articles on AI and the future of work. But too often, this discourse focuses solely on job losses, coupled with sensational headlines. While technologies continue to evolve rapidly, meaning that we can’t yet drew firm conclusions, it is also true that we need to look at the labour market effects much more carefully. This entails understanding how AI impacts not only the quantity of jobs but also the quality of them and the nature of work, in terms of wages, working conditions and rights more broadly. The larger issue of inequality is also crucial.  

2. In your view, what are the current and potential future impacts of automation on developing economies? 

Currently, the impact of digitalization and AI in developing economies is less than in advanced economies, primarily due to the digital divide and differences in the structure of economies. While the possible negative effects on jobs are less (but not zero), we also know, as shown by ILO research, that the digital divide prevents developing countries from taking advantage of the benefits from new technologies, such as AI.  

But there are two caveats to this view that developing countries are not being affected by AI. First, digital tools and platforms are growing rapidly, even if the use of generative AI is more confined to small part of the population – see the rise of digital payment systems, such as M-PESA in Eastern Africa, and the rapid emergence of digital labour platforms, both location-based and online, which are creating opportunities and challenges in all countries. Second, emerging technologies, such as AI, continue to evolve so we cannot assume that the situation today will hold in the coming months and years. We need to continue to monitor the situation.  

3. How do these impacts influence groups in vulnerable situations, such as women, youth, and migrants? 

A key lesson from centuries of technological change is that there are both winners and losers as economies and labour markets adjust. ILO research has shown that women are more susceptible to the automation effects of AI due to their overrepresentation in occupations that are most exposed, such as administrative jobs. Recent data on actual labour market trends (as opposed to potential effects) is telling us that we should be worried about how generative AI is impacting young people – evidence is emerging to suggest that there is a more negative effect of AI on young labour market entrants in such countries as the US. At the same time, there are range of use cases that can help certain groups access new learning and employment opportunities. For example, under the  PROSPECTS programme, the ILO is supporting young people in remote areas of Kenya through digital skills training and mentoring programmes to access online job opportunities. 

4. While we talk about the quantity of jobs impacted by AI, what would you say about the quality of jobs impacted? 

As already mentioned, we need to focus on not only implications of AI for job quantity but quality as well, which is where the largest effects are likely to emerge. AI impacts tasks and won’t eliminate most jobs entirely. But these changes can lead to effects on wages, depending on how demand for the occupation shifts, and working conditions due to the impact of AI in the workplace.  

We already see the use of algorithmic management tools, which are getting supercharged by AI, for recruitment, allocating tasks, monitoring and evaluating workers. While this has the potential to improve productivity, it poses challenges in terms of workers’ agency and the nature of their jobs. Key is transparency on the use of these tools, matched by dialogue to ensure that new technologies can be beneficial to both the enterprise and their workers.  

5. There’s often a regional divide in how AI is deployed and its benefits are distributed. How can we bridge that gap—both nationally and globally? 

A digital divide exists both within and between countries due to differences in access to digital infrastructure and skills. The share of the population using the Internet in 2024 reached 93 per cent in high-income countries compared with just 27 per cent of the population in low-income economies. Even in advanced economies, such as in the European Union, access to AI is uneven with higher rates of adoption in richer countries and larger firms. Within countries, access to broadband (optic fibre) Internet and training programmes is more limited in rural areas. In response, greater investments are needed in infrastructure and skilling to ensure that these gaps are reduced. Support is also needed to ensure that there are opportunities for developing economies to build their own AI ecosystems in terms of both development and deployment of new technologies in different languages and adapted to country-specific contexts.   

The share of the population using the Internet in 2024 reached 93 per cent in high-income countries compared with just 27 per cent of the population in low-income economies.

6. What kind of policy frameworks do we need to ensure that AI benefits all workers? 

From an ILO perspective, the Decent Work Agenda remains key for assessing both the benefits and challenges arising from the development and deployment of AI. In practice, responding to the opportunities and challenges posed by AI will involve applying existing policies and regulations, while adapting and developing new strategies and governance frameworks where needed, in line with international labour standards and through social dialogue (e.g., to address the platform economy).  

There are three areas we need to look at: first, address the negative impact of AI through redeployment, social protection and active labour market policies (e.g., employment services); second, enhance digital skilling and upskilling to support access to new technologies, along with measures to assist small businesses to overcome the digital divide and take advantage of opportunities; and third, strengthen governance mechanisms to ensure rights are protected in the workplace (e.g., safeguarding against discriminatory algorithms). 

7. If you could deliver just one message to global policymakers about AI and employment, what would it be? 

AI is creating both new opportunities and challenges in the world of work and to ensure that the benefits are broadly shared, we need to assess the impact of AI on both the quality and quantity of jobs, and respond through employment policies and other measures, backed by the latest evidence and social dialogue.  

Prepared by Paolo Alberto Baudino, Oscar Schwartz Blicke and Maurizio Michael Habib

Published as part of the Financial Stability Review, November 2025.

Hedge funds represent a relatively small segment of the euro area investment fund sector and comprise both AIF and UCITS hedge funds. The total assets of euro area hedge funds stood at around €660 billion in the third quarter of 2025, equivalent to roughly 3% of the investment fund sector’s total assets. In the EU, hedge funds may fall either under the Alternative Investment Fund Managers Directive (AIFMD)^[1] or the Undertakings for Collective Investment in Transferable Securities (UCITS) Directive.^[2] AIF hedge funds are usually marketed to wealthy investors and are predominantly held by euro area investment funds. They offer limited liquidity, by allowing redemptions only quarterly or even annually (often with advance notice), for example, and by imposing lock-up periods on initial investments. By contrast, UCITS hedge funds are more accessible to retail investors and other non-bank sectors – with euro area households and insurance corporations each holding around 15% of the shares in such funds. As these funds often allow investors to redeem shares on a high-frequency basis, the sector is more exposed to fund share redemptions during market turmoil. UCITS hedge funds account for about 30% of the overall hedge fund sector in terms of shares issued (Chart A, panel a) as well as total assets.

Chart A

UCITS hedge funds exhibit higher retail participation and use derivatives more intensively than do AIF hedge funds

a) Investor base and shares issued by euro area hedge funds, by fund type	b) Euro area hedge funds’ financial leverage and gross derivatives exposure, by fund type	c) Gross derivatives exposures of euro area UCITS hedge funds and derivative type distribution
(Q3 2025; percentages, € billions)	(Q3 2025; total assets divided by shares issued, derivative gross notional divided by shares issued)	(Q3 2025, derivative gross notional divided by shares issued)

As UCITS hedge funds have relatively high derivatives exposure and leverage, they warrant attention from a financial stability perspective. Both UCITS and AIF hedge funds employ a wide range of investment strategies, including leveraged trades, to achieve positive absolute returns. Because of regulatory constraints on borrowings,^[4] UCITS hedge funds make less use of financial leverage than AIF hedge funds do, with a total assets/equity ratio of 1.3 for UCITS hedge funds versus 1.7 for AIF hedge funds. However, synthetic leverage through derivatives is more pronounced in UCITS hedge funds, with gross notional derivatives exposure reaching up to 12 times equity for fund categories such as global macro strategies (Chart A, panels b and c).^[5] In addition, UCITS hedge funds hold a lower proportion of highly liquid assets (e.g. cash and sovereign bonds) than AIF hedge funds do.^[6] This leaves them more vulnerable to liquidity risk from redemption shocks or margin calls. Although some research has been carried out on the performance of UCITS hedge funds, this box sheds light on their liquidity and leverage-related risks, given their importance for financial stability.^[7]

Procyclical flows and larger redemptions from leveraged funds in times of stress can lead to asset sales and mounting liquidity pressures during periods of high market volatility. Evidence from a panel of 457 UCITS hedge funds shows that their flows are procyclical, positively correlated with past returns (Chart B, panel a) and in line with the findings for other fund categories.^[8] Although the analysis does not indicate that leverage generally amplifies the flow procyclicality of UCITS hedge funds, it does show larger outflows from leveraged UCITS hedge funds in periods of market stress (Chart B, panel b). Since fund share redemptions may force funds to sell assets when markets are under pressure, leveraged funds could be required to close larger positions, thereby amplifying stress.

The use of derivatives by UCITS hedge funds can intensify liquidity pressures via margin calls. Derivatives positions, which can be used for hedging or for leverage, are subject to margin requirements. During periods of elevated price volatility and significantly negative returns, margin calls on these derivatives positions tend to increase (Chart B, panel c), further straining a fund’s liquidity.^[9] This exacerbates the challenges faced by leveraged UCITS hedge funds, as they have to manage liquidity to meet both margin calls and redemption requests simultaneously. Interaction between these factors can heighten liquidity strains and contribute to broader market stress under adverse market conditions.^[10]

Chart B

Flows into UCITS hedge funds tend to be procyclical, while margin calls may intensify liquidity risk

a) Average fund-level flows into euro area UCITS hedge funds, by lagged return level	b) Average fund-level flows into euro area UCITS hedge funds, by synthetic gross leverage level	c) Average fund-level daily posted variation margin of euro area UCITS hedge funds, by negative return level
(Jan. 2019-Oct. 2025; standardised values, percentages)	(Jan. 2019-Oct. 2025; standardised values, log of derivative gross notional as a percentage of TNA)	(Jan. 2020-Oct. 2025; percentages of TNA, percentages)

A robust stress-testing framework for leveraged UCITS hedge funds is essential to ensure their resilience and limit the risks to financial stability in turbulent market conditions. The combination of outflows and margin calls on derivatives positions can intensify liquidity pressures for UCITS hedge funds during periods of stress. This raises concerns about the ability of such funds to manage the challenges and contributes to broader financial instability. These dynamics highlight the need for strengthened risk management and comprehensive stress-testing practices to safeguard financial stability during episodes of market turmoil.

Finally, authorities should be equipped with suitable tools to limit excessive leverage in UCITS hedge funds and mitigate the build-up of risks during periods of market stress. While authorities have tools that enable them to contain excessive leverage in AIFMD-compliant funds, they do not have such tools for UCITS hedge funds. The Eurosystem suggests introducing discretionary powers that would allow authorities to impose stricter leverage limits on these funds when they pose risks to financial stability.^[12] It also recommends that all UCITS hedge funds should be required to report their leverage using the commitment approach.

Prepared by Maciej Grodzicki, Urtė Kalinauskaitė, Benjamin Klaus, Chloe Larkou, Francesca Lenoci and Allegra Pietsch

Published as part of the Financial Stability Review, November 2025.

The US dollar assets and liabilities of euro area banks arise mainly from their capital market activities. Capital market activities are often characterised by short maturities and require daily marking to market and margining. These features may pose a liquidity risk to banks in the event of abrupt market movements. US dollar funding and hedging instruments provided by banks are also important for euro area corporates and non-bank financial institutions, especially when exchange rates move rapidly. This box presents the business rationale for euro area banks’ US dollar activities and aims to assess the associated financial stability risks.

US dollar activities are concentrated among euro area global systemically important banks (G-SIBs), which intermediate US dollars to other European parties. In contrast to other major currencies, US dollar activities relate almost exclusively to wholesale business.^[1] The breakdown of euro area banks’ US dollar assets and liabilities reveals the high weight of capital market activities relative to loans and deposits from the non-financial sector (Chart A, panel a). Other financial assets and liabilities, which include primarily the positive fair value of derivatives, are the largest balance sheet position denominated in US dollars. They are followed by repo borrowing, debt securities funding, holdings of debt securities and deposits taken from banks and other financial institutions.

Banks’ dollar-denominated credit exposures are largely limited to holding high-quality debt securities and lending to the non-financial corporate sector. Euro area banks’ debt securities holdings denominated in US dollars consist mainly of US Treasuries and agency mortgage-backed securities, followed by debt issued by non-US governments and financial institutions. These securities qualify as high-quality liquid assets. Euro area banks’ dollar-denominated lending is estimated to be close to €700 billion at the least (9.2% of the total loan book), with most of this going to non-euro area corporate and non-bank financial clients (Chart A, panel b).^[2]

Chart A

Capital markets business dominates in euro area banks’ US dollar activities

a) Dollar-denominated assets and liabilities of 48 euro area significant institutions, by bank type	b) Dollar-denominated securities holdings and bank lending of euro area banks
(Q4 2024, € trillions)	(Q4 2024, percentages)

The US dollar activities of euro area banks in capital markets represent a diverse set of financial services to the economy. Euro area banks, especially some of the G-SIBs, are present in US money markets, where they act as intermediaries by sourcing funding from money market funds and lending the proceeds to hedge funds on a secured basis.^[3] Euro area investment funds, life insurers and pension funds invest in dollar-denominated assets, despite their euro-denominated obligations to fund-shareholders or policyholders. Euro area banks facilitate these counterparties’ needs to mitigate the resulting currency risk by engaging in FX swaps, effectively receiving US dollars and paying euro to investment funds, insurance corporations and pension funds. Euro area banks partially hedge this currency risk by taking opposite positions with global banks (Chart B, panel a). These US dollar liabilities are not visible on bank balance sheets.^[4] Euro area banks also provide currency hedges to euro area exporters and importers, although such hedging trades are on a smaller scale than those associated with euro area financial investors. For instance, as of July 2025, banks are facilitating US dollar payments to non-financial corporations via currency swaps, primarily to stabilise such corporations’ import costs rather than to manage dollar-denominated revenue flows.

While asset-liability mismatch appears to be limited in extent, banks are nonetheless taking liquidity risk due to mismatches between counterparties providing and receiving funding. Some banks mitigate liquidity risks further by running maturity-matched repo books and do not use volatile short-term repos to fund illiquid long-term assets. However, these strategies do not fully eliminate liquidity risk. While banks hold sizeable quantities of high-quality liquid assets, dollar outflows in an extreme scenario could exhaust their capacity to raise cash through repos, FX swaps and the sale of such assets.^[5] The value of liquid assets may also decline in these circumstances, exacerbating liquidity pressures.^[6] Although net outflows of US dollars could be covered by US dollar liquid assets in the long term, the net outflows are concentrated in the short term. Some banks may require additional funding in US dollars or rely on inflows of US dollars from maturing short-term assets to remain liquid during financial stress (Chart B, panel b). However, collecting these cash inflows would imply that they reduce US dollar funding to counterparties.

Chart B

Euro area banks’ US dollar intermediation activities

a) Euro area banks’ net dollar-denominated FX swap and CIRS positions	b) Cumulative contractual gap between US dollar inflows and outflows to/from euro area banks
(Q2 2025, € billions)	(Q4 2024, percentages of US dollar HQLA)

Maintaining adequate balance sheet capacity is necessary to enable banks to act as shock absorbers. If euro area banks reduce their dollar intermediation, their counterparties could face difficulties funding or hedging dollar-denominated investments and may need to sell such assets. Capital and US dollar liquidity buffers provide the balance sheet space required by banks to offer financial services in US dollars to their counterparties in times of financial stress. Capital headroom could be needed to absorb the increase in capital requirements associated with higher currency volatility and counterparty credit risk. Although liquidity risk may not materialise in banks’ own balance sheets and there is no regulatory requirement for banks to match the currencies of liquid assets to the currencies of liabilities, banks should hold liquid US dollar assets to counterbalance outflows and act as a stabilising intermediary.