# Hungarian DPA fines Budapest Bank HUF 250m for AI emotion analysis of customer-service calls Source: https://www.bankingnewsai.com/ai-incidents/budapest-bank-2022-emotion-ai-call-analysis-fine Last updated: Oct 21, 2026 Budapest Bank Zrt. · Governance finding · severity 2 of 5 In a decision dated 8 February 2022 (NAIH-85-3/2022), Hungary's data-protection authority, NAIH, fined Budapest Bank HUF 250 million, about €634,000, for automatically analysing recorded customer-service calls with AI-based speech-signal processing that detected callers' and employees' emotions and keywords and ranked dissatisfied customers for call-backs. Per NAIH, the bank, which introduced the software in 2017, gave customers no proper notice of the voice analysis, offered no workable right to object and relied on a legitimate-interest assessment that ignored proportionality, breaching several GDPR articles. NAIH ordered the bank to stop analysing emotions. The bank had told NAIH the software contained no AI and made no automated decisions; NAIH pointed to the bank's own impact assessment, which said the recordings were analysed automatically with AI. ## Key facts | | | | --- | --- | | Date | Feb 8, 2022 | | Institution | Budapest Bank Zrt. | | Country | Hungary | | AI system | AI-based speech-signal processing that detected emotions, keywords and silences in recorded customer-service calls and ranked callers for call-backs | | AI? | Described as AI by the institution or regulator — The institution, a regulator or a court called the system AI. | | Category | Regulatory finding (Governance finding) | | Status | Regulator finding | | Severity | 2 of 5 — A regulatory finding with a HUF 250m fine (about €634,000, under $1m) and an order to stop emotion analysis; the processing touched 1 to 1.5 million recorded calls a year, but no financial loss to customers was found. | ## What happened? According to the decision, Budapest Bank introduced the voice-analysis software on 26 May 2017 to make the work of about 180 telephone staff more effective. The bank told NAIH that the software, using "mesterséges intelligencián alapuló beszédjel-feldolgozás" (AI-based speech-signal processing), recognised waiting, silence and people talking over each other, found keywords from a dictionary the bank had built to filter out complaining customers and prevent churn, and detected "érzelmi / hangulati elemeket" (emotional or mood elements) in the recordings. The analysis covered 1 to 1.5 million calls a year. [1] NAIH opened an ex officio procedure on 22 September 2021 after an earlier investigation triggered by a customer complaint; the procedure covered the period from May 2018 to September 2021. It found that the bank used the results to decide which dissatisfied customers to call back, and analysed the emotional state of both the caller and the call-centre employee. The bank's privacy notice mentioned the analysis only in general terms, with no material information about the voice analysis itself, and according to the bank's own statements it had for years not given callers proper notice or a right to object because it had decided it was unable to. [1][2][3] The bank told NAIH that "A Szoftver mesterséges intelligenciát nem tartalmaz, automatizált döntést nem hoz" ("The Software does not contain artificial intelligence and makes no automated decision"; our translation). NAIH set this against the bank's own impact-assessment document, which stated that the recordings "mesterséges intelligencia segítségével automatizáltan kerülnek elemzésre" ("are analysed automatically with the help of artificial intelligence") and that the data were suitable for profiling and scoring. NAIH held that the bank should have assessed, before starting AI-based automatic voice analysis, whether it could be run in compliance with data-protection rules. [1] NAIH also questioned whether the system worked: per an attachment the bank supplied, no emotion was recognisable in 91.96% of cases. It found breaches of GDPR Articles 5(1)(a) and (b), 6(1), 6(4), 12(1), 13, 21(1) and (2), 24(1) and 25(1) and (2), and said emotion-based voice analysis could lawfully rest only on freely given, informed consent. [1][2] ## Timeline - May 26, 2017 — Budapest Bank introduces the voice-analysis software in its telephone customer service. [1] - Jul 5, 2021 — In a preliminary investigation prompted by a complaint, the bank answers NAIH's questions about the voice analysis. [1] - Sep 22, 2021 — NAIH opens an ex officio data-protection procedure covering processing since 25 May 2018. [1] - Feb 8, 2022 — NAIH issues decision NAIH-85-3/2022: a HUF 250m fine and an order to stop analysing emotions. [1][4] - Apr 2022 — DLA Piper reports the decision as a record GDPR fine by NAIH for the unlawful use of artificial intelligence. [2] ## What was the impact? - Financial: HUF 250,000,000 fine (about €634,000 per the CMS GDPR Enforcement Tracker) - Customers: Recorded calls of customers and other callers, about 1 to 1.5 million a year (81,500 calls a month in 2021), were analysed; call-centre employees' voices were analysed too. [1][4] ## How did Budapest Bank Zrt. respond? No public statement by Budapest Bank found as of 2026-10-21 (we searched Hungarian and English press and legal coverage). In the proceedings the bank argued that the processing rested on its legitimate interests in retaining customers and improving efficiency, that the system stored no identifying data, and that the software contained no AI and made no automated decisions, its results being usable only through human interpretation. [1][3] ## What was the outcome? NAIH fined the bank HUF 250 million and ordered it to change the processing so that emotions are no longer analysed, to give proper notice and a right to object, and to limit analysis of employees to what is needed, proving compliance within 60 days or proving that the processing had stopped. The decision could be challenged in an administrative court; we found no report of a court challenge. [1] ## FAQ ### What did Budapest Bank's AI do with customer calls? Per NAIH's decision, the software analysed recorded customer-service calls with AI-based speech-signal processing: it detected silences and overlapping speech, searched for keywords from a bank-built dictionary, and detected emotional or mood elements in callers' and employees' voices, then ranked calls so staff could call back dissatisfied customers. ### Was the system really AI? The bank told NAIH its software contained no artificial intelligence and made no automated decisions. NAIH relied on the bank's own impact assessment, which said the recordings were analysed automatically with the help of artificial intelligence, and on the vendor platform's description of machine learning and neural networks. The regulator treated it as AI. ### How much was Budapest Bank fined, and what else did NAIH order? NAIH imposed a HUF 250 million fine, about €634,000 according to the CMS GDPR Enforcement Tracker, in its decision of 8 February 2022. It also ordered the bank to stop analysing emotions, to give proper notice and a right to object, and to prove compliance within 60 days. ### Did customers lose money? No financial loss was found. The harm NAIH identified was to privacy rights: roughly 1 to 1.5 million calls a year were analysed for emotions without proper information, a valid legal basis or a workable right to object, and the analysis fed decisions about which customers to call back. ## Sources 1. [NAIH-85-3/2022 határozat (Budapest Bank Zrt., hangfelvétel-elemzés)](https://naih.hu/hatarozatok-vegzesek?download=517:mesterseges-intelligencia-alkalmazasanak-adatvedelmi-kerdesei) — Nemzeti Adatvédelmi és Információszabadság Hatóság (NAIH), Feb 8, 2022 (official): "a Hatóság hivatalból utasítja az Ügyfelet, hogy akként módosítsa adatkezelési gyakorlatát, hogy az megfeleljen az általános adatvédelmi rendeletnek, azaz a hangelemzés során az érzelmeket ne elemezze" 2. [Hungary: Record GDPR fine by the Hungarian Data Protection Authority for the unlawful use of artificial intelligence](https://privacymatters.dlapiper.com/2022/04/hungary-record-gdpr-fine-by-the-hungarian-data-protection-authority-for-the-unlawful-use-of-artificial-intelligence/) — DLA Piper (Privacy Matters), Apr 12, 2022 (news): "The bank used the results of the analysis to determine which customers should be called back by analysing the emotional state of the caller using an artificial intelligence-based speech signal processing software" 3. [Budapest Bank stung with fine for carrying out emotional AI analysis incorrectly](https://www.lewissilkin.com/en/insights/2022/05/19/budapest-bank-stung-with-fine-for-carrying-out-emotional-ai-analysis-incorrectly-102hohn) — Lewis Silkin, May 19, 2022 (news): "Despite the bank stating that the systems used did not store any identifiable personal data, the Hungarian DPA held that this was not the case" 4. [GDPR fine against Budapest Bank Zrt. (Hungary)](https://www.enforcementtracker.com/ETid-1244) — CMS GDPR Enforcement Tracker, Feb 8, 2022 (incident index): "The Hungarian DPA (NAIH) has fined Budapest Bank Zrt. EUR 634,000." All AI incidents at banks: https://www.bankingnewsai.com/ai-incidents --- Canonical page: https://www.bankingnewsai.com/ai-incidents/budapest-bank-2022-emotion-ai-call-analysis-fine Part of [BankingNewsAI](https://www.bankingnewsai.com/) — the register of AI incidents at banks: https://www.bankingnewsai.com/ai-incidents. Markdown versions of every reference page: append `.md` to the page URL; index at https://www.bankingnewsai.com/llms.txt.