Article 14 of the Proposed EU KIDS Act: A New Relational-Safety Layer in EU Digital Law

 

Theodoros Zarkos, MSc, Medical Ethicist - Social Scientist - European Social Policy and Social Rights; Independent Researcher, Kiato, Greece

Photo credit: maido155, via Wikimedia Commons

 

The European Commission’s proposal for an EU KIDS Act, published on 17 September 2026, contains a provision that is easy to misread if it is viewed in isolation. Article 14 would create specific duties for AI companions and general conversational chatbots accessible to minors. Yet the proposal does not begin from an empty legal field. Online platforms, AI systems and the processing of children’s personal data are already governed by overlapping EU rules. The legal question is therefore not whether Article 14 “regulates AI for minors” for the first time, but what kind of legal layer it adds to the Digital Services Act (DSA), the Commission’s 2025 minors Guidelines, the AI Act and the GDPR.

Primary text: COM(2026) 681 final - proposed EU KIDS Act

From a horizontal child-safety duty to AI-specific obligations

The DSA already establishes a broad child-protection baseline. Article 28(1) requires providers of online platforms accessible to minors to put in place appropriate and proportionate measures ensuring a high level of privacy, safety and security for minors on their service. The provision is deliberately general. It does not attempt to enumerate every risky interface or emerging technology.

See: Digital Services Act, Article 28

The Commission’s 2025 Guidelines under Article 28(4) made that baseline more concrete. For AI features integrated into an online platform accessible to minors, the Guidelines recommend that the feature should not be activated automatically, that minors should not be encouraged or enticed to use it, that it should be easy to turn off, and that providers should assess risks before making it available. They also recommend visible warnings distinguishing AI interaction from human interaction. The Guidelines are non-binding; the Commission nevertheless uses them as a reference point when assessing compliance with Article 28(1).

See: Commission Guidelines on the protection of minors under the DSA (2025)

Article 14(2) of the proposed KIDS Act closely tracks part of that earlier guidance. Where an AI companion or general conversational chatbot is embedded in a social network, video-sharing service or online game, the host provider would have to ensure that the feature is not activated automatically or displayed prominently, that minors are not encouraged to use it, and that they can opt out easily at any time. In that limited deployment context, the proposal would move several safeguards from Commission guidance into an operative legislative rule. In doctrinal terms, this operates largely as specification of an existing DSA child-safety direction rather than the creation of an entirely new policy idea.

The new element is relational design itself

Other parts of Article 14 go further than the DSA guidance. Article 14(1)(a) would require providers to avoid design features and system behaviours that simulate interpersonal relations where they are likely to create emotional dependencies. It also imports the anti-addictive and excessive-use measures set out elsewhere in the proposal. The legal object is therefore not merely “harmful content” or “screen time”; it is a design pathway in which interpersonal simulation and engagement architecture may shape a minor’s continued use of the system.

The AI Act does not make this provision redundant. Article 5 of the AI Act prohibits certain manipulative or deceptive AI practices, and the exploitation of vulnerabilities including those related to age, when the statutory conditions concerning behavioural distortion and harm are met. Article 50, separately, requires transparency where natural persons interact directly with AI, unless the artificial nature of the interaction is obvious in the circumstances. These are important rules, but they perform different functions. Article 14(1)(a) would create a child-specific, ex ante design obligation directed at relational dependency. It does not depend on first establishing the full set of conditions required by the AI Act’s prohibited-practices regime.

See: Artificial Intelligence Act, Articles 5 and 50

Conceptually, the proposed text does not prohibit anthropomorphism, emotional tone or perceived closeness as such. Its trigger is narrower: interpersonal simulation that is likely to create emotional dependency. That wording leaves a substantial implementation question about how “likely” and “emotional dependency” will be operationalised in testing, supervision and enforcement.

What the empirical evidence does—and does not—show

An adolescent–parent experiment published online in September 2026 compared matched written chatbot replies using either a relational or a more transparent conversational profile. In 284 dyads, adolescents rated the relational profile more highly for such immediate qualities as anthropomorphism, emotional closeness and trust. Preference for that profile was associated with lower family-relationship quality and higher reported stress and anxiety. The experiment supports a design-sensitivity claim: conversational framing can affect immediate perceptions. It does not establish that chatbot use caused those family or mental-health circumstances, or that long-term dependence follows (Kim, Xie and Yang, 2026).

The evidentiary boundary is equally important in regulatory design. Self-selected adolescent narratives of overreliance and disengagement illuminate possible mechanisms, but cannot establish their frequency or a causal effect. In addition, children’s evolving capacities require age-sensitive safeguards rather than an assumption that maximal parental control is always desirable. An Article 14 assessment should therefore distinguish experimentally observed responses, plausible longer-term risks and independently justified rights protections (Namvarpour et al., 2026; UN Convention on the Rights of the Child, Articles 5 and 12).

Research source: Kim, Xie and Yang (2026)

Conversational memory: overlap with the GDPR, but a different legal function

Article 14(1)(b) introduces another distinctive rule. By default, a covered system should not use information or analysis derived from a minor’s prior interactions in subsequent interactions, subject to the exceptions set out in the provision. This is more precise than saying that the proposal simply “bans memory”. It establishes a default configuration concerning the reuse of prior-interaction information and analysis.

The GDPR is obviously relevant whenever conversational memory involves personal data. Its principles of lawfulness, fairness and transparency, purpose limitation and data minimisation continue to apply, as does Article 25 on data protection by design and by default. Article 8 addresses a child’s consent in relation to information society services where processing relies on consent, with Member States permitted to set the relevant age between 13 and 16. None of those provisions, however, is a general relational-safety rule for conversational systems.

See: General Data Protection Regulation, Articles 5, 8 and 25

The proposed KIDS Act operates at a different legal level. The GDPR asks whether and how personal data may lawfully be processed. Article 14(1)(b) would additionally regulate how a product for minors is configured to carry relational context from one interaction into the next. In practice, the two regimes may apply to the same technical feature, but they answer different legal questions.

Testing and monitoring create an ex ante / ex post cycle

Article 14 also proposes a governance cycle that is more specific than the horizontal baseline. Before placing a covered system on the market or putting it into service, providers would have to carry out state-of-the-art evaluation and testing for risks to minors’ health, safety, fundamental rights and physical, mental and emotional well-being and development, and implement appropriate safeguards. After market entry, providers would have to monitor harms and emerging risks and respond to serious incidents involving minors. The post-market monitoring duty contains a narrow exception for micro and small enterprises; it is not a general exemption from Article 14.

For children below 13, Article 14(1)(d) would also require access to be enabled and controlled through the guardian tools established by Article 20. That rule should not be confused with GDPR Article 8. One concerns access architecture and guardian controls under the proposed KIDS Act; the other concerns the validity of consent as a lawful basis for specified personal-data processing.

Existing enforcement is not evidence of developmental efficacy

An earlier enforcement example helps separate existing obligations from the proposal’s added design duties. In its decision of 10 April 2025 concerning Replika, the Italian data-protection authority examined transparency, legal bases and deficient age-assurance measures for a conversational service. The case shows that regulators can already intervene in access and data-governance arrangements involving minors without Article 14. It does not establish that relational chatbots cause long-term developmental harm or demonstrate that the proposed relational safeguards will be effective. Article 14 would extend the architecture of prevention; its practical effects would still require evaluation.

Primary enforcement decision: Italian Garante, 10 April 2025  

What is specification, what is extension, and what remains mixed?

My recent evidence-to-regulation study approached this layering question by separating three categories: Specification, where Article 14 largely turns an existing baseline into a more concrete rule; Extension, where it introduces a materially new child-specific relational safeguard; and Mixed, where both functions are present. This is an analytical classification, not language used by the Commission itself.

On that basis, the activation and disengagement controls in Article 14(2) are the clearest case of Specification because they closely translate the Commission’s 2025 guidance for integrated AI features into proposed operative obligations. The dependency-oriented design rule, the prior-interaction default, and the treatment of human-relationship displacement are better understood as Extensions of the identified baseline. Addictive or compulsive engagement and trust or emotional closeness have a mixed relationship with earlier instruments because relevant duties already exist, but Article 14 reorganises them around a specifically relational and child-focused context.

The same study also cautions against treating legal correspondence as proof of regulatory effectiveness. Its scientific corpus comprised 36 active sources and 111 mechanism-level extraction rows, alongside 20 normative and legal sources. Nine studies met a strict criterion for direct minor-specific empirical evidence. No long-term causal harm study was identified within that strict subset. The evidence was especially limited for conversational memory as an isolated causal mechanism, and much of the empirical literature remains weighted towards adolescents rather than younger children.

Underlying study: Bioethical Governance of Relational AI for Minors: Evidence-Regulation Alignment in Article 14 of the Proposed EU KIDS Act

The limits of layering

The proposal’s layered structure also creates boundaries that should remain visible. Article 14(2) applies its activation, prominence, encouragement and opt-out controls only when the qualifying AI is deployed as a functionality of a social network, video-sharing service or online game. The underlying relational mechanisms can also arise in standalone systems, but those specific paragraph 2 controls do not follow them there. Other Article 14 duties do apply more broadly to covered AI companions and general conversational chatbots.

Likewise, trust and emotional closeness are not independently regulated outcomes. They may be relevant to dependency, transparency, testing and monitoring, but the proposal does not create a general legal threshold for when a minor’s trust in a system becomes impermissible. Nor does the proposal itself establish that a particular safeguard will reduce long-term dependency or displacement of human relationships. Those are implementation and evidence questions that arise after the legal duty has been defined.

A relational-safety layer, not a replacement regime

Article 14 is therefore better read as a new relational-safety layer within an already dense EU digital framework. The DSA continues to supply the horizontal platform duty; the 2025 Guidelines remain relevant to the interpretation and practical application of that duty; the AI Act retains its own thresholds for prohibited manipulation, exploitation and transparency; and the GDPR continues to govern personal-data processing. The proposed KIDS Act adds rules that are more directly tied to the architecture of interaction between minors and conversational AI.

The resulting structure is cumulative rather than substitutive. Some provisions would harden existing Commission guidance into legislation. Others would extend the baseline by addressing emotional dependency, prior-interaction reuse, pre-market testing and post-market monitoring in a child-specific way. The overlap is real, but so are the differences in legal trigger, scope and function. If the proposal advances through the ordinary legislative procedure, those distinctions will matter when lawmakers, providers and regulators decide what Article 14 requires in practice.

*Prior related work and methods disclosure. This post develops a distinct doctrinal analysis drawing on the author’s DOI-registered, non-peer-reviewed preprint and related research synthesized in an unpublished monograph. A related, broader evidence-focused commentary by the author appeared in The European Scientist in September 2026. Editorial assistance with structure and wording involved an AI tool; the author is responsible for source verification, interpretation and the submitted text.

When Does Credit Stop Being Finance? Debt Collection before the CJEU

 

Cătălin Gabriel Stănescu, Reader in Law at Newcastle University and member of the European Commission’s Financial Services Users Group.

Photo credit: BNP Paribas HQ, by Boubloub, via Wikimedia Commons

A pending reference from the Greek Council of State brings an old question back before the Court of Justice: when credit is transferred to a purchaser and its administration and recovery are entrusted to a separate company, how should EU financial law classify the activities of that company? In Trapeza tis Ellados II, C-210/26, the Court is asked whether an undertaking engaged exclusively in credit servicing, without assuming credit risk, qualifies as a financial institution under the Regulation (EU) No 575/2013 on prudential requirements for credit institutions (Capital Requirements Regulation, CRR).

In a 2021 article,  I asked whether informal debt collection qualified as a financial service under EU law and what that qualification would mean from a legal perspective (pp 535-538). The answer was less straightforward than the question suggested. Treating collection as a financial service might activate additional protection for those subjected to it, but it might also make that protection depend on the origin of the debt. The difficulty itself was older: my 2015 analysis of MKG-Kraftfahrzeuge-Factoring had already shown how factoring, debt purchasing and collection could overlap without becoming identical.

Two recent CJEU judgments make the pending reference particularly interesting. In Kosmiro, the Court describes debt collection, for VAT purposes, as financial transactions designed to obtain payment (para 50). In IFIS NPL INVESTING, it excludes the purchase or management of non-performing loans from the economic activities covered by a particular definition of financial institution (para 37). The reference in Trapeza II concerns a servicer whose activities include administering and collecting debts, negotiating with borrowers and restructuring repayment obligations. The question is not simply whether these judgments can be reconciled, but how far a classification reached under one EU regulatory regime can be carried over into another. I argue that regulatory differentiation does not require conceptual separation. Lending, assignment, servicing and collection may attract different legal classifications while remaining connected stages in the life of the same credit relationship.

1.     The question in Trapeza II

The reference in Trapeza II was lodged by the Greek Council of State on 13 March 2026 and is currently pending. It arises from sanctions imposed by the Bank of Greece on the former managing director of a credit-servicing company (summary of the request, paras 6-8). The company was authorised and supervised by the Bank but operated exclusively as a servicer. Its activities included legal and accounting monitoring, recovery, negotiations with debtors, settlements and debt restructuring (paras 14-15). It assumed no credit risk (para 29).

The question is whether such a company is a “financial institution” under Article 4(1)(26) of CRR. The definition applicable at the material time covered an undertaking, other than an institution or pure industrial holding company, whose principal activity was acquiring holdings or pursuing one or more activities listed in points 2-12 and 15 of Annex I to Directive 2013/36/EU on access to the activity of credit institutions and the prudential supervision of credit institutions (Capital Requirements Directive, CRD). The dispute turns on two possible Annex I classifications: point 2, ‘lending’, which expressly includes factoring with or without recourse, and point 11, ‘portfolio management and advice’.

The Bank of Greece characterises the work as administrative factoring and, where securitised loans are concerned, portfolio management (summary of the request, para 10). The referring court provisionally rejects both propositions. That conclusion forms part of a broader line of reasoning. Drawing on Lineas – Concessões de Transportes and Others, joined Cases C-207/22, C-267/22 and C-290/22, judgment of 26 October 2023, the referring court starts from the proposition that an undertaking which does not carry out activities in the financial sector cannot qualify as a financial institution (para 23). It then reads the Annex I activities as, in principle, activities linked to the banking or financial sector. It invokes Directive (EU) 2021/2167 on credit servicers and credit purchasers as confirmation, in its view, that credit servicers are not financial institutions within the meaning of the CRR and do not simultaneously carry out activities in the financial sector (para 27). It ultimately concludes that companies performing only the specified servicing activities, without assuming credit risk, ‘do not carry out activities in the banking or financial services sectors’ (para 29). The classification is of immediate importance because it determines whether the sanctions applicable to financial institutions could be used against the applicant. The reference nevertheless raises a broader issue: how EU financial law classifies an undertaking that performs functions arising from a credit relationship while neither originating the credit nor assuming the associated credit risk.

There is also a temporal limit. Regulation (EU) 2024/1623 (CRR III) amended that definition with effect from 1 January 2025. The current Article 4(1)(26) also expressly includes an “ancillary services undertaking” as an alternative basis for classification as a financial institution. Article 4(1)(18), as amended, defines such an undertaking by reference to a principal activity consisting, inter alia, of a direct extension of banking or other activities ancillary to banking. That express route to financial-institution status was absent from the version of Article 4(1)(26) applicable to the dispute. The judgment will therefore interpret a superseded definition. Its reasoning on factoring, portfolio management and credit servicing may nevertheless remain relevant under current law, even though the amended definition now provides an additional route to financial-institution status.

2.     Factoring as debt collection

Kosmiro, C-232/24, judgment of 23 October 2025, concerned two factoring models. Under trade factoring, the factor purchased invoices and assumed the debtors’ default risk. Under invoice factoring, the debts remained with the client, which retained the default risk while obtaining finance secured by those debts. In both arrangements, the factor administered and collected the debts (paras 11-13). The VAT question was whether the relevant fees constituted consideration for exempt credit or taxable debt collection under Article 135(1)(b) and (d) of Directive 2006/112/EC on the common system of value added tax (VAT Directive).

The Court began from the debt-collection exception to the exemption for transactions concerning debts. Exemptions are interpreted strictly, while the exception, which restores taxation, is interpreted broadly (paras 46, 49). Debt collection refers to “financial transactions designed to obtain payment of a pecuniary debt” (para 50). That concept encompasses both factoring models (paras 51-53).

That conclusion builds on an established line of VAT-related activities case law. In MKG-Kraftfahrzeuge-Factoring, C-305/01, judgment of 26 June 2003, the Court held that factoring is merely a variant of the more general concept of debt collection, whatever the manner in which it is carried out (paras 76-77). In AXA UK, C-175/09, judgment of 28 October 2010, it added that debt collection need not concern an overdue debt or involve coercion (paras 31-35). The resulting VAT concept is therefore not confined to coercive recovery of debts already in default.

In Kosmiro, the Court held that, in principle, factoring constitutes a single economic transaction whose main purpose is to transfer responsibility for recovery and collection from the client to the factor and splitting that transaction would be artificial (para 60). In trade factoring, the funds paid for the purchased debts were not a loan to the client and there was consequently no credit relationship between factor and client (para 61). In invoice factoring, by contrast, financing was provided, but not independently of the debt-collection service, of which it was the corollary (paras 62-63). The Court distinguished arrangements whose essential purpose was independently the provision of financing (paras 64-66). Importantly, it rejected the argument that all financing arrangements must receive the same VAT treatment: the legislative distinction between debt collection and the granting of credit permits “financing arrangements subject to VAT and financing arrangements exempt from VAT” to coexist (para 67). The relevant invoice-factoring fees therefore constituted consideration for a single and indivisible debt-collection service subject to VAT (paras 68-69). 

There are firm limits. The Court classifies the transactions for VAT purposes but does not define the institutional perimeter of prudential supervision. It establishes that factoring falls within debt collection, not that every collector performs factoring. Nor does the debtor purchase the service examined in Kosmiro: the factor supplies it to the creditor-client. Finally, purchasing debts need not supply a remunerated service at all. In GFKL Financial Services, C-93/10, judgment of 27 October 2011, a price reflecting the actual economic value of defaulted claims was not consideration for a service to the seller (paras 22-26).

Kosmiro consequently provides no trans-substantive definition of financial services. Its narrower significance is nevertheless important. EU VAT law describes debt collection as consisting of financial transactions designed to obtain payment of a pecuniary debt, and Kosmiro demonstrates that financing and collection can form part of the same economic arrangement without becoming identical legal services. What it undermines is therefore any assumption that the functions of financing and collection must always be legally or economically separate.

3.     The classification puzzle

The difficulty is that the Court has recently used a different regulatory definition to reach a conclusion that appears to leave little room for treating the purchase or management of non-performing loans as a financial activity. In IFIS NPL INVESTING, C-65/25, judgment of 11 June 2026, the Court stated that the economic activities covered by Directive (EU) 2015/849 on the prevention of the use of the financial system for the purposes of money laundering or terrorist financing (Fourth Anti-Money Laundering Directive, AMLD) do not include the purchase or management of non-performing loans (para 37). The conclusion is relevant for Trapeza II because the Greek undertaking did not purchase the claims it serviced, but its activities included their administration, collection, negotiation and restructuring. The question is whether the exclusion of NPL “management” in IFIS also encompasses those distinct servicing functions.

IFIS arose from Italian legislation governing bulk assignments of non-performing loans during the period before the expiry of the deadline for transposing Directive 2021/2167. The referring court asked, among other things, whether entities carrying out such assignments had to be subject to anti-money-laundering supervision.

In this instance, the definition before the Court was Article 3(2)(a) of AMLD, not Article 4(1)(26) CRR. It covered an undertaking other than a credit institution which carried out one or more activities listed in points 2-12, 14 and 15 of CRD Annex I. That statutory route nevertheless matters for Trapeza II because it incorporates the two Annex I activities at issue there: point 2, which includes factoring, and point 11, portfolio management and advice. Yet the Court does not analyse either Annex I activity separately. Its unqualified exclusion of NPL purchase and management consequently raises the question whether either activity can bring an undertaking within the corresponding CRR definition.

That passage is difficult to confine to its facts. Unlike the historical CRR definition, the AMLD did not require an Annex I activity to constitute the undertaking’s principal activity. If the purchase or management of NPLs does not fall within the relevant Annex I activities at all, making that activity the undertaking’s principal business cannot by itself bring it within the CRR definition. IFIS therefore points strongly towards exclusion if the Greek servicer’s work is characterised simply as the management of NPLs.

IFIS bears directly on Trapeza II, but does not resolve it. The dispute in IFIS arose from bulk assignments of NPL portfolios, whereas the Greek undertaking in Trapeza II neither acquired the claims it serviced nor assumed their credit risk. Paragraph 37 of the IFIS judgment extends the Court’s conclusion beyond the purchase of NPLs to their “management”, but does not examine what that activity encompasses. In particular, the Court did not consider whether the administration, collection, negotiation and restructuring performed by a credit servicer could constitute factoring under point 2 of Annex I to the CRD, or whether the management of a portfolio of credit claims could fall within “portfolio management and advice” under point 11 of that Annex. Instead, the reasoning moves directly from the scope of the AMLD to the exclusion of NPL purchase and management. The temporal analysis in IFIS does not qualify that conclusion. Paragraphs 31–36 of the judgment explain why Directive 2021/2167 did not apply to the transfers at issue and could not be applied retroactively; paragraph 37 states a separate conclusion concerning the scope of the AMLD. That conclusion must therefore be confronted on its own terms.

The two definitions also perform different functions. The AMLD identifies obliged entities for the purposes of anti-money-laundering supervision, whereas the CRR definition operates within a system of prudential regulation and financial-sector supervision. In Lineas, the Court interpreted the historical CRR definition in that regulatory context. It concluded that an undertaking whose sole activity consisted in acquiring holdings in companies which did not carry out activities in the financial sector was not a “financial institution”. Its reasoning emphasised the distinction drawn by the CRR between holdings in financial-sector entities and holdings outside that sector, and the inconsistency that would result from subjecting the latter to rules designed for the former (paras 69-81).

Those different regulatory purposes matter, but they do not by themselves resolve Trapeza II. Both definitions incorporate points 2 and 11 of the same CRD Annex I. If those common terms produce different classifications in the AMLD and CRR contexts, the difference must be explained through the wording, structure and purpose of the respective regimes. Regulatory purpose alone cannot determine what counts as factoring or portfolio management.

Trapeza II thus offers the Court an opportunity to provide the analysis compressed into paragraph 37 of IFIS. It could explain whether servicing without credit risk can constitute factoring under point 2; what “portfolio management and advice” means in point 11 when the portfolio consists of credit claims; and whether the servicer’s combination of collection, negotiation and restructuring affects either classification. If the result remains exclusion, the reasoning should identify why those activities fall outside the particular definition rather than turn that exclusion into a general proposition that servicing and collection are external to finance.

Consumer law reveals the relevance of that distinction. In Gelvora, C-357/16, judgment of 20 July 2017, the Court held that the concept of commercial practices directly connected with the sale of a product extends beyond the conclusion of the contract to its performance, to include measures taken to obtain payment (paras 19-21). The claims at issue originated in consumer credit, and their recovery could fall within the Unfair Commercial Practices Directive even after assignment to a debt collection agency (paras 22-25). Excluding credit-repayment transactions following assignment could undermine consumer protection by encouraging traders to separate the recovery phase in order to escape the Directive (para 28). The Directive could also apply where unilateral collection measures were pursued in parallel with formal enforcement (paras 29-31).

Gelvora does not classify debt collection as a financial service. Paragraph 25 appears to distinguish the two activities: the collection agency does not provide consumer credit “as such”. Its significance lies instead in the continuity recognised for the particular purposes of consumer protection. Assignment changes the creditor and recovery changes the activity, but neither necessarily breaks the legal connection between collection and the credit from which the payment obligation arose.

INKO, C-127/15, judgment of 8 December 2016, provides a complementary, activity-specific example. Under Directive 2008/48/EC on credit agreements for consumers, a debt collection agency acting on behalf of a lender to conclude a rescheduling agreement for unpaid credit must be categorised as a “credit intermediary” within the meaning of Article 3(f) (paras 42-44). The classification follows not from debt collection alone but from the additional function performed by the agency within the existing credit relationship: concluding the rescheduling agreement on the creditor’s behalf. The Court further held that acting as a credit intermediary only in an ancillary capacity may remove particular pre-contractual information obligations without altering that classification (paras 47-53).

4.     Continuity without regulatory identity

The case law does not produce a single answer to whether collection is a “financial service”. And, perhaps, it should not. Lending creates credit; purchasing transfers a claim; servicing manages rights and obligations arising from it; collection seeks payment; and enforcement invokes legally compulsory procedures. Those functions may overlap, but they are not identical and may justify different authorisation, conduct, prudential and tax rules.

The difficulty begins when a conclusion designed for one of those purposes is detached from it. Under Directive 2021/2167, a credit purchaser does not require authorisation to operate because it acquires existing non-performing credit rather than creating new credit and raises negligible systemic risk (Recital 40). IFIS excludes the purchase or management of NPLs from the economic activities relevant to the AMLD, while Trapeza II asks whether particular servicing functions fall within the historical CRR definition. A particular acquisition of debts may not constitute a remunerated service for VAT purposes. None of those situations establishes, without another legal step, that collection has become external to the financial relationship from which the payment obligation arose.

Directive 2021/2167 embodies this distinction. A credit servicer manages and enforces rights and obligations related to a creditor’s rights under a non-performing credit agreement, or to the agreement itself, on behalf of a credit purchaser. Credit-servicing activities include collecting or recovering payments, specified renegotiation, complaints administration and informing borrowers of changes in interest rates or charges or of payments due (Article 3(8) and (9)(a)-(d)). Rather than simply designating servicers as credit institutions or financial institutions, the Directive subjects them to a bespoke authorisation and supervisory regime (Articles 4-14 and 21-23).

At the same time, the regulatory object remains connected to the underlying credit relationship. Article 10 of Directive 2021/2167 regulates the conduct of purchasers and servicers towards borrowers, requiring good faith, fairness and professionalism and prohibiting harassment, coercion and undue influence. It also requires specified information to be communicated after transfer and always before the first debt collection. The limited prudential concerns associated with credit purchasers do not, however, displace the conduct obligations applicable to them and their servicers. Yet Union and national consumer-protection rules continue to apply and borrowers retain the rights arising from the initial credit agreement. Regulatory identity changes; legal continuity remains.

This also separates different forms of risk. Whether an undertaking assumes the economic loss resulting from non-payment may be highly relevant to prudential classification. It says much less about the conduct risks created by controlling communications with borrowers, repayment negotiations and collection. A servicer’s absence of credit risk may therefore matter to one regulatory classification without establishing that its activities are unrelated to the underlying financial relationship. Undertakings might pose little or no systemic risk, but that does not mean no risk to consumer-debtors. Directive 2021/2167 responds to such conduct risks through authorisation and conduct obligations tailored to servicing rather than lending.

It should be emphasized that the Directive is selective evidence, not a complete solution. Its framework concerns non-performing credit agreements issued by credit institutions established in the Union and contains significant exclusions (Articles 1-2). It does not create a common regime for the collection of rent, utility debts or other non-credit claims. That limitation leaves intact a problem I identified in my earlier work: if protection against abusive collection depends on the regulatory identity of the debt or collector, comparable collection practices may attract different legal responses for reasons having little to do with the risks those practices create.

The answer is not to declare every collector a financial institution, but to resist allowing institutional definitions to determine questions they were not designed to answer. Prudential classification should follow prudential purposes. Consumer-facing collection should remain subject to effective conduct controls because of the activity and its connection to the underlying obligation. Where gaps remain outside the sectoral regime, they require direct regulation rather than conceptual disappearance.

Credit does not stop being finance at a single identifiable moment. Default, assignment and outsourcing redistribute its functions and risks. They may change the applicable service, institution and regulatory regime without erasing the relationship between the claim being collected and the transaction that created it. IFIS requires purpose-specific exclusions to be taken seriously. Kosmiro, Gelvora and INKO require equal caution before turning such an exclusion into a general separation between finance and collection. Trapeza II is an opportunity to articulate both propositions and to clarify a boundary that EU law has so far defined only imperfectly.

Two Regimes of Recovery: Greece’s Exit from Macroeconomic Surveillance and the Social Findings That Do Not Close a File

 Joanna Demopoulou*

 

Photo credit: Mstyslav Chernov, via Wikimedia Commons

 

* Joanna Demopoulou holds a PhD in International Affairs and is a former Executive in Residence at the Geneva Centre for Security Policy (GCSP). Her research examines the legal and institutional architecture of public authority and governance, focusing on how law structures, distributes, and constrains public authority across European and international institutions.

 

 

On live television in late September, Greece’s deputy foreign minister, Haris Theocharis, told a fellow panelist that his party represents “noikokyrides”, householders who manage their affairs responsibly, while opposition parties represent the “kakomoiris”, the unfortunate. He later called the phrase a mistake. The remark is domestic party rhetoric, not the language of any EU institution, and this post does not treat the Union as its author. It is a symptom of an institutional environment that makes it easy, more than a decade after Greece entered its first adjustment programme, to recode continuing hardship as a private failing. That environment is the subject here.

 

An asymmetry, not just two numbers

 

The familiar observation is that macroeconomic recovery and social recovery are different things, and that a country can score well on one while scoring badly on the other. That is true but not new, and by itself it is a complaint about incomplete data, not a constitutional argument. The sharper claim is about institutional force: what each kind of finding is allowed to do. Some of the Union’s indicators can close a file. Others can only describe what remains in it.

 

Under the Macroeconomic Imbalance Procedure, a Member State is placed into one of several formal categories, and movement between them triggers or lifts surveillance, and in the worst case, correction. A finding of “imbalances” carries institutional consequences; it is a status, not a description. A finding of “no imbalances” removes that status and the surveillance that comes with it. Continuing hardship, measured by Eurostat’s indicator of poverty and social exclusion or by the Commission’s own Social Convergence Framework, is real institutional knowledge, produced and published by the same Union. It has, as far as I have found, no comparable status that opens or closes. It accumulates as description.

 

One answer is that this is simply a difference of subject matter: debt and deficits are quantities a procedure can act on directly, while poverty has many causes and no single lever. That is a reason to design a different kind of instrument, not a reason for the Union to have no closing instrument at all. The Social Convergence Framework already collects comparable indicators and already names “overall challenges” for specific countries. What it does not produce, on the evidence available, is a formal status with an identifiable condition for exit. The asymmetry is better described as a feature of institutional design than as a fact of nature about the difference between fiscal and social policy; nothing here requires assuming that anyone chose it deliberately.

 

Closing the macroeconomic file

 

The 2026 European Semester Spring Package (COM(2026) 200), adopted on 3 June, found that Greece, the Netherlands and Sweden “are assessed as no longer experiencing imbalances”. For Greece this followed a finding of “imbalances” in the 2025 cycle. Alongside the Spring Package, a separate Post-Programme Surveillance assessment (C(2026) 4000), whose stated objective is to confirm that a country “maintains its capacity to service its debt”, concluded of each of Ireland, Greece, Cyprus and Portugal in turn that it “retains the capacity to service its debt”. Together these findings close the imbalance status that had attached to Greece and record that, on the Union’s own assessment, it retains the capacity to service its debt.

 

The same spring, Eurostat reported that 27.5% of people in Greece were at risk of poverty or social exclusion in 2025, the second-highest share in the Union after Bulgaria, against an EU average of 20.9%. The Commission’s Social Convergence Framework, published weeks earlier, named Greece among six Member States facing “overall challenges to upward social convergence”, alongside Bulgaria, Italy, Latvia, Lithuania and Romania. Both are Union findings, produced through the same European Semester as the imbalance procedure, by the same Commission, in the same year. Neither carries a formal status with an identifiable condition for exit. There is no announced moment at which a Member State is declared to have exited social hardship the way it exits an imbalance. No scoreboard entry reads “no longer experiencing social exclusion.”

 

A pattern with a history

 

The pattern predates this recovery. It was present at the start of the adjustment programmes, and admitted along the way. The European Parliament’s 2014 resolution on the employment and social aspects of the Troika found that Greece’s programmes rested on “a mistaken assumption regarding the economic multiplier effect, resulting in failure to act in time to protect the most vulnerable against poverty”, and it regretted that “proper impact assessments were not made of their distributional impact on different groups of society”. Parliament called on the Commission to take social indicators into account in any renegotiation of the programmes.

 

The Commission’s own ex-post evaluation, published nine years later in 2023, concluded that the Greek programmes “did not initially perform as expected, but that their results improved over time and their main objectives have been achieved”. The objectives referred to are fiscal and financial: the evaluation’s own executive summary names “fiscal consolidation to ensure debt sustainability” and Greece’s gradual return to sovereign markets as the programmes’ main objectives. The 2023 evaluation neither revisits the 2014 finding about households left unprotected against poverty nor contradicts it. That kind of finding is simply not what a programme evaluation is built to reopen. The two documents, nine years apart, describe the same programmes from within two different regimes of assessment, and only one of those regimes produces a verdict of success.

 

Why this is a constitutional question

 

It would be possible to read all of this as a policy complaint dressed in legal language: the Union should care more about poverty, full stop. The claim here is different. Article 9 TFEU and Article 3(3) TEU are not external moral commentary on economic governance; they are commitments inside the same legal order that produces the Semester, the imbalance procedure and the surveillance reports. This post does not argue that the Treaties require a social counterpart to the imbalance procedure, or that the absence of one is unlawful in itself. The question is why one part of that legal order can close a file while the other, under the same Treaties, can only describe.

 

What the constitutional texts actually commit to

 

The Treaties are not silent on this asymmetry, even if they do not resolve it. Article 9 TFEU requires the Union, in defining and implementing its policies, to take into account “a high level of employment, the guarantee of adequate social protection, the fight against social exclusion”. Article 3(3) TEU sets out objectives of the Union rather than a requirement of that kind: “a highly competitive social market economy, aiming at full employment and social progress”, and the combating of “social exclusion and discrimination”. Neither provision creates a directly enforceable right, and neither did the work in this recovery that the imbalance procedure did.

The Charter of Fundamental Rights binds Member States only when they implement Union law. Article 1 on human dignity and Article 34 on social security therefore cannot be read as a freestanding review of a minister’s remark, or of Greek welfare policy generally. But the Union’s institutions were not outside the Charter during the adjustment era. In Ledra Advertising, the Court of Justice held that the Charter is addressed to the Commission even when it acts within the framework of the European Stability Mechanism, and that the Commission, under Article 17(1) TEU and Article 13(3) and (4) of the ESM Treaty, must ensure that a memorandum of understanding is consistent with the fundamental rights the Charter guarantees. That reasoning concerned Cyprus, but the same provision of the ESM Treaty governed Greece’s third programme, financed through the ESM from 2015. The Commission negotiating that programme’s conditionality was, on the Court’s own logic, not acting in a Charter-free space. The constitutional commitments were therefore legally live while the programme took its social shape. What they did not generate, then or since, was a closing mechanism comparable to the one that, a decade later, closed the macroeconomic file. A Member State’s exit from the imbalance procedure is announced by name, in a document with legal consequences. I have found no act that records how the requirements listed in Article 9 TFEU, which the Union must take into account in defining and implementing its policies, were weighed in reaching that finding.

 

From institutional story to personal blame

 

None of this explains a minister’s choice of words; it describes the space in which that choice becomes intelligible. When the dominant institutional narrative is that a country has recovered, continuing difficulty is harder to read as an unfinished collective problem and easier to read as an individual one. If the country succeeded, the reasoning runs, whoever is still struggling did something wrong, or belongs to the group the recovery was never going to reach. “Noikokyris” and “kakomoiris” give that reasoning a name; they do not invent it. Pressed on air by an opposition MP, Tonia Antoniou, Theocharis said his party represents both the “noikokyrides” and people who “get by with difficulty”, which kept the division intact even as he tried to soften it.

 

“Recovered” is not a loose figure of speech picked up by commentators. It is the closest available paraphrase of what the Semester’s own category change says about Greece, a status built on indicators designed to answer exactly that question. Against that background, any remaining hardship needs a separate explanation, and “noikokyris/kakomoiris” supplies one.

 

This is a diagnosis about institutional design, not an accusation against any single actor. Those who speak of recovery are not wrong about what they measure: the debt ratio has fallen, and Greece’s return to market financing is real. The claim is narrower. An architecture that lets one kind of finding close a file and confines another to permanent description makes it easier for political language to treat the second as somebody’s personal failing. Nothing here shows that the architecture causes that language; it shows what it leaves open for that language to do.

 

Several Union reports on poverty already exist, so another one is not the remedy. The task, instead, is to ask, as an open constitutional question rather than a ready design, what it would take for the Union’s own commitments in Article 9 TFEU and Article 3(3) TEU to carry institutional weight of a kind comparable to the imbalance procedure’s, and what would have to be true before a Member State’s file could be said to reflect both. This post does not answer that question; it argues that the gap in weight, not merely in numbers, is where the answer has to start. Until that changes, the Union will keep producing two kinds of knowledge about the same country, and the difference between them lies less in what the Union knows than in what each kind is institutionally permitted to do: one can end a crisis, and the other can only keep describing it.