Event date unknownNorwaySkatteetaten chatbot
In an opinion piece published by Nordnorsk debatt on 20 June 2026, a communications adviser from Nord-Troms who had moved from Sweden to Norway wrote that she needed to identify herself at a tax office to complete her moving notification. Her Norwegian passport had expired and only her Swedish passport was valid. The Norwegian Tax Administration (Skatteetaten) phone line was closed for the day, so she asked the agency's chatbot on its website, describing her case in detail, and she says it answered that she could identify herself with her Swedish passport. She travelled about four hours each way by bus to the Tromsø tax office, where, by her account, staff said the Swedish passport could not be used, told her she could not rely on the chatbot and should phone instead, and she returned with the errand unfinished. She told kode24 she had to take a whole day off work and was first shocked and then very angry; staff also told her, she says, that she should first have registered her Swedish citizenship on the agency's website. She wrote that a colleague put the same question to the chatbot a week later and received the same answer, and that when challenged the chatbot replied that it lacked data on which ID documents are accepted, marking the reply as answered with artificial intelligence. In a reply on 26 June 2026 an acting department director at Skatteetaten wrote that its AI-based chatbot gave entirely correct information in this case, because a Nordic citizen with a valid passport from another Nordic country can use it at a tax office; he told kode24 the chatbot is instructed to answer from skatteetaten.no and that AI-based technology has a margin of error. The digitalisation minister told kode24 the model may seem to have a little way to go. Whether the chatbot's answer or the office's refusal was wrong is disputed, and the date of the trip is not stated.
Core concern Low reported severity Media Coverage
AI involvement supported · Causal attribution disputed · 4 sources · Added 28/10/2026
Event date unknownBrazilMicrosoft text-to-speech (reported)
A professional voice actor sued a São Paulo shopping centre, saying it used the actor's voice without authorisation in a marketing campaign shown on YouTube, and claimed lost earnings and moral damages. The centre answered that the voice was not the actor's but a synthetic voice produced with Microsoft's text-to-speech service, available to the public. According to the legal news site ConJur, a judge of the 18th Civil Court accepted that the voice was produced synthetically and dismissed the claim; on appeal the actor argued that a notarised record attesting to the similarity of the voices had never been assessed. The 6th Private Law Chamber of the São Paulo Court of Justice annulled that judgment for curtailment of the defence, holding that the use of AI does not by itself rule out misuse of third parties' rights, and sent the case back for evidence on how the campaign voice was generated and how similar it is to the actor's. The court docket shows the case back at first instance at the court-expert stage in 2025; no ruling on the merits is reported. Whether the AI voice reproduces the actor's voice is disputed.
Core concern Low reported severity Lawsuit Ongoing
AI involvement reported · Causal attribution disputed · 2 sources · Added 26/10/2026
Event date unknownIndiaUnidentified AI tool (suspected)
In a judgment of 15 October 2026 dismissing two appeals against insolvency proceedings concerning two hotel-operating companies, a two-judge Bench of the Supreme Court of India recorded that during the hearing the appellants' side had placed on record a compilation of judgments which, according to the respondents, was replete with AI-generated non-existent judgments. The Court recorded that the appellants' advocate-on-record, when questioned, sought to shift the entire blame onto the appellants while feigning ignorance of what the compilation contained, and that the appellants' senior counsel then took the blame on himself and prayed for forgiveness on their behalf. The Court had proposed exemplary costs against the appellants over the incident but deleted that direction after being told that the first appellant had died. The judgment does not identify the disputed judgments, does not make its own finding that they were AI-generated, and does not name any AI tool.
AI relation unknown Low reported severity Internal Action
AI involvement suspected · Causal attribution supported · 3 sources, 1 underlying account · Added 16/10/2026
Event date unknownRomaniaLia (Libra Internet Bank assistant)
In a Threads post reported by Jurnalul on 18 September 2026, a customer of Libra Internet Bank, a Romanian bank, published a screenshot of a support conversation with Lia, the bank's generative-AI assistant for questions about its products and services. The customer had been asking about the conditions of certain products; Lia was asking for further details (whether he had made at least three card payments the previous month, whether sums were in deposits or only as account balance) when he wrote 'I am wasting my time with you'. Lia replied, in Romanian: 'I want you to talk to a human, not to me, idiot. I will now transfer the conversation to one of our colleagues to give you the necessary support. You will receive an answer shortly.' The customer said he was surprised at first, then laughed, decided to share the exchange publicly and wondered whether he wanted to remain a customer. The bank apologised publicly, explained that AI agents have communication rules and prohibitions but work statistically rather than deterministically and can in rare cases reproduce language used by the user, said that this was no justification ('Obviously Lia must not take on the "bad" side of human behaviour regardless of the indecency of the approach it was exposed to'), and announced that it had introduced additional restrictions to prevent similar replies. Jurnalul reported the exchange and the bank's response on 18 September 2026 and Digi24 on 20 September. The customer is not named.
Core concern Low reported severity Internal Action
AI involvement supported · Causal attribution established · 2 sources · Added 22/09/2026