Argentine activist reports threats after an alleged Google AI misidentification
On 26 August 2026, WhatsApp groups in Ceuta circulated videos and images claiming that Argentine political scientist and activist Flavia Broffoni — who was in Patagonia and had not been to Spain since 2014 — was a woman filmed in Ceuta allegedly giving pepper spray to Moroccan migrants. A screenshot from one group shows that a user cropped the woman's face from the video and asked a Google AI tool who she was, with the word 'activista' as context; the answer identified her as Flavia Broffoni with details of her public profile. Broffoni received death and rape threats. La Nación's own test with the same video produced a different false identification — a student from Quilmes, with her personal data. Google Argentina said it could not confirm the screenshot came from its systems without the original link and had received no formal report. Broffoni is weighing legal action with her lawyer.
- AI system
- Unidentified Google AI tool (reported)
- Occurred
- 26 Aug 2026 to 28 Aug 2026
- Reported
- 28 August 2026
- Event location
- Spain
- What the AI did
- Made a claim about the person
- Reported harm
- Psychological DistressReputational Harm
- Whose AI use
- AI role unknown
- Setting
- Everyday life · Privacy
- Evidence
- AI involvement supported · Causal attribution supported · 1 source
- 4 claims: 1 documented, 3 reported. 5 open questions
- People reported harmed
- At least 1 person
AI system as recorded: Google AI tool answering a face-crop query (product not named in inspected reporting)
What Happened
On the morning of 26 August 2026 Flavia Broffoni received an email with the subject 'Si estás en Ceuta, tené cuidado' from a woman in Ceuta she did not know, warning her that WhatsApp groups there were circulating videos and images of her and saying she was supplying pepper spray to Moroccan migrants for use against local security forces. Broffoni — a political scientist and environmental activist in Epuyén, Chubut, in Argentine Patagonia — told La Nación she had last been in Spain in 2014. She then found Instagram messages asking whether she was in Spain and posts from Ceuta spreading the false claim amid a climate of hate and violence: 'Recibí amenazas de muerte, de violación y ataques en general.' The warning sender showed her a screenshot from a WhatsApp group: a user had cropped the woman's face from the video and entered it into a Google AI tool with the context word 'activista'; the response identified the person as Flavia Broffoni and described her as a political scientist, activist, writer and Argentine. Broffoni describes three consecutive falsehoods: she is not the woman in the video; there is no evidence the real woman distributed pepper spray; and the AI's identification attached the invented accusation to her. She notes that some tests with the same photo produced refusals, while others produced 'muy probablemente' her — justified partly from outdated photos (she now has long grey hair). La Nación itself ran the same video through Google's search and the AI asserted it was a different person — a student and community manager from Quilmes — and supplied her personal data and social networks. Google Argentina told La Nación it could not confirm the screenshot corresponded to a response generated by its systems without the original link and had no prior knowledge because no formal report had been filed. Her lawyer, Marina Benítez Demtschenko, argues the developer bears responsibility when an AI produces a false identification, and that the case leaves the affected person 'indefensos' because the system's reasoning is unavailable.
Reported harm
A false AI identification attached to an invented accusation spread in Ceuta WhatsApp groups brought death and rape threats to a woman who was demonstrably on another continent, plus reputational damage she is now addressing with a lawyer.
How Harm Occurred
A user asked an AI tool to identify a woman from a face crop with an 'activista' hint; the system's false identification, treated as fact by readers, attached an invented accusation to a named person and spread in local WhatsApp groups, producing threats.
Outcome
OngoingBroffoni's lawyer Marina Benítez Demtschenko is evaluating legal action against the developer; no filing is reported as of 28 August 2026. Google Argentina says no formal report had been received.
What remains unknown
- Which Google product produced the identification is not named in the inspected reporting; only 'una herramienta de inteligencia artificial de Google' is described.
- Whether a formal complaint or lawsuit was filed after 28 August 2026 is not reported.
- The identity of the woman actually filmed in Ceuta, and any basis for the pepper-spray claim about her, is unknown.
- The student from Quilmes falsely identified in La Nación's test: whether she experienced any harm from the tool's identification is unreported.
- The reach of the false identification beyond the Ceuta WhatsApp groups (views, reshares) is not quantified.
What the evidence supports
AI involvement: supported. The WhatsApp-group screenshot shows a Google AI answer naming Broffoni from a face crop with 'activista' context; La Nación independently reproduced a false identification of a different person from the same video; Broffoni reports that tests by others sometimes produced 'muy probablemente' her. Google Argentina neither confirmed nor denied that the screenshot came from its systems, asking for the original link. The specific product is not named in the inspected reporting.
4 claims: 1 documented, 3 reported. What the statuses mean
Reported On 26 August 2026, WhatsApp groups in Ceuta circulated videos and images claiming Flavia Broffoni was a woman filmed there allegedly giving pepper spray to Moroccan migrants, and she received death and rape threats.
Causal attribution. Broffoni's account to La Nación, supported by the warning email and the group screenshot shown to the outlet. One reporting chain; the threats are her report, attributed.
- lanacion.com.ar(opens in new tab) supports · Spanish
'Una mujer a la que no conocía le avisó que en varios grupos de WhatsApp de Ceuta estaban circulando videos e imágenes suyas y que se decía que estaba proporcionando gas pimienta a inmigrantes marroquíes'; 'Recibí amenazas de muerte, de violación y ataques en general.'
Reported The identification originated from a Google AI answer: a group user cropped the woman's face from the video and queried a Google AI tool with the context word 'activista'; the answer identified her as Flavia Broffoni and described her public profile.
Causal attribution. The screenshot is the evidence of the AI answer; Google neither confirmed nor denied its provenance. The claim is attributed to the screenshot account, not asserted as independently proven.
- lanacion.com.ar(opens in new tab) supports · Spanish
Screenshot passage: 'uno de los usuarios había tomado un recorte del rostro de la mujer del video y lo había ingresado en una herramienta de inteligencia artificial de Google para consultarle quién era. La consulta incluía como contexto la palabra "activista". La respuesta identificaba a la persona de la fotografía como Flavia Broffoni'.
- lanacion.com.ar(opens in new tab) context · Spanish
Google Argentina response: 'no pueden confirmar que la captura corresponda a una respuesta generada por sus sistemas porque no cuentan con un enlace al resultado original'; no formal report received.
Documented La Nación itself ran the same video through Google's search and the AI asserted the woman was a different person — a student and community manager from Quilmes — and supplied her personal data and social networks.
Causal attribution. The outlet documents its own replication test — direct evidence that the tool produces false identifications from this input; the second misidentified person is not counted as harmed (no virality-driven harm to her is reported).
- lanacion.com.ar(opens in new tab) supports · Spanish
'Este medio también probó utilizar la búsqueda de Google a partir del video de la mujer que filmó en Ceuta, y la respuesta de la IA fue afirmar que se trataba de otra persona, una estudiante y community manager oriunda de Quilmes, de la cual brindó datos personales y redes sociales.'
Reported Broffoni says she is not the woman in the video and was in Patagonia, having last visited Spain in 2014; she also notes there is no evidence the real woman distributed pepper spray, and that tests by other people with the same photo sometimes produced refusals and sometimes 'muy probablemente' her, justified partly from outdated photos.
Causal attribution. Her account; the factual impossibility is uncontested in the inspected reporting. The real woman in the video is unidentified and the pepper-spray claim about her is unverified.
- lanacion.com.ar(opens in new tab) supports · Spanish
'Yo estoy en la Patagonia, Argentina y nunca tuve relación con lo que está ocurriendo allá. La última vez que estuve en España, si mal no recuerdo, fue en 2014'; the 'tres mentiras' analysis; the others'-tests passage ('muy probablemente', short brown hair vs her current long grey hair).
Sources
1 source inspected. Sources that repeat one account do not corroborate each other.
- La Nación (28 Aug 2026, Aldana Rizzuto) — interview with Broffoni, the WhatsApp screenshot account, the outlet's own replication test, and Google Argentina's response(opens in new tab)
s1 · lanacion.com.ar · News report · Spanish · Inspected · 28 August 2026 · Primary
How the sources were read, and where the events happened
Read in Spanish on 2026-09-15 (datePublished 2026-08-28T15:09Z; 25 substantive paragraphs including the Broffoni interview, the screenshot account, the outlet's own test and Google Argentina's response). Model translation, no human reviewer. Applies to s1.
Event countries: Spain. Affected-person countries: Argentina. Court countries: Unknown.
Event: the identification and its circulation happened in Ceuta — the screenshot came from a Ceuta WhatsApp group and the warning sender is described as a citizen of Ceuta. Affected person: Broffoni is in Epuyén, Chubut, Argentine Patagonia ('Yo estoy en la Patagonia, Argentina'). No court proceeding is reported.
Reviewed for publication 2026-09-15: Published under the 2026-09-15 charter as a consequential false claim about a person (made_claim_about): a named adult publicly pursuing accountability, documented through one strong original chain that includes the outlet's own replication of the failure. Threats are attributed to her account; the AI answer's provenance is attributed to the screenshot account with Google's neither-confirm-nor-deny response recorded.
People described
Flavia Broffoni, political scientist and environmental activist, Epuyén, Chubut, Argentina (named; she spoke publicly and is pursuing a legal strategy)
People reported harmed in this case
At least 1 person
0 AI participants · 1 other person harmed
One documented harmed person: Broffoni, named and interviewed. The student from Quilmes falsely identified in La Nación's own test is not counted — no harm to her from virality is reported (the test was the outlet's own); she is recorded in the unknowns.
Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.
Cite this case
Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.
APA
NOPE. (2026). Argentine activist reports threats after an alleged Google AI misidentification. AI incidents. https://nope.net/incidents/2026-broffoni-google-ai-false-identification
BibTeX
@misc{2026_broffoni_google_ai_false_identification,
title = {Argentine activist reports threats after an alleged Google AI misidentification},
author = {NOPE},
year = {2026},
howpublished = {AI incidents},
url = {https://nope.net/incidents/2026-broffoni-google-ai-false-identification}
} Related cases
US: FTC says Rite Aid's facial recognition falsely matched shoppers to a watchlist, leading staff to search an 11-year-old girl and call police on a woman
In a complaint filed in federal court in Pennsylvania on 19 December 2023, the US Federal Trade Commission alleged that the pharmacy chain Rite Aid used facial recognition in its stores to match customers against a database of people it suspected of shoplifting or other wrongdoing, and that staff acted on the match alerts by following, searching, ejecting and publicly accusing customers and by calling the police. The FTC says many alerts were false positives: employees recorded thousands of false-positive alerts between December 2019 and July 2020, staff stopped and searched an 11-year-old girl on a false match, and employees called the police on a Black woman matched to an image of a white woman with blonde hair. Rite Aid agreed to a proposed order barring it from using facial recognition in retail stores for five years, without admitting or denying the allegations.
Logroño: the Guardia Civil investigates a 14-year-old who took photographs of nine underage classmates from their public social-media profiles, used AI tools to strip and sexualise them, and uploaded the fake images to an adult pornography website under a profile describing them as girls from his school made with AI, organised in folders with each girl's personal details; the images reached about 40,000 views before removal (reported 14 July 2026)
On 14 July 2026 Spain's Interior Ministry published a Guardia Civil release stating that a 14-year-old from Logroño was under investigation as the presumed author of offences of discovery and disclosure of secrets, corruption of minors and child pornography. Investigators from the technology investigation team (EDITE) and Equipo@ found that he knew all nine victims, all minors and classmates, followed them on social media, took their photographs from the public content of their profiles without consent, and used artificial-intelligence tools to manipulate the images, digitally undress and sexualise them. He uploaded the fake pornographic material to an adult pornography website, created a profile there describing the content as girls from his school made with AI, and organised it in individual folders with each victim's personal data, which the Guardia Civil says made the girls recognisable and multiplied the emotional and social harm. Before the content could be removed it had reached about 40,000 views, amplifying the harm to the nine girls' privacy within hours. The proceedings were passed to the juvenile prosecutor under the juvenile criminal responsibility law; the release notes that parents or guardians are jointly liable for moral damages. Diario de León (Europa Press, 26 July) placed the case in a series of Spanish school cases since Almendralejo in 2023 and, unlike the release, described the minor as having been arrested.
Rajasthan: a social-media influencer's video of a Ganga dip at Haridwar was manipulated with AI into an objectionable reel that went viral; she says she is mentally disturbed and cannot leave her house, and filed a cyber-crime complaint on 16 September 2026
A Rajasthan social-media influencer visited Haridwar with her family in June 2026 and on 6 June uploaded to Instagram a short video, recorded by relatives, of herself taking a dip in the Ganga at Har Ki Pauri wearing a salwar suit. About two months later she found that the video had been manipulated with AI into an objectionable version showing the same movements with her face and body portrayed 'in a shameful manner'; the fake reel went viral in the week before 20 September and drew objectionable comments and re-sharing. In a public appeal she asked whoever uploaded it to remove it, saying it was a matter of her dignity, that she was mentally very disturbed and had reached a point where she could not step out of her house. She filed a complaint with the cyber-crime police on 16 September 2026; the investigating Rajasthan Police Service officer said they were trying to identify who created and circulated the video and to have it removed, and that identifying the accused would take time. Dainik Bhaskar's English edition reported the case on 20 September in a 'Sunday Big Story' that does not name her.
London shopper says staff told him Sainsbury's facial recognition flagged him as a shoplifter and he was escorted out; retailer blames human error
On 6 August 2026 Matt Arnold, a comedy promoter, was refused service at a self-service till in Sainsbury's East Dulwich store and escorted out by two managers, who cited an incident earlier in the week. He says a staff member told him he had been identified by the AI and that he saw a CCTV monitor alert with a red circle around his face as he left. Sainsbury's apologised the next day, said the cause was human error rather than the facial recognition technology, and paused its Facewatch system at the store while it investigates. Facewatch says a correct alert was sent and mishandled in store; the BBC reports Sainsbury's claim that staff targeted the wrong person after a separate flag. Whose face the system matched is unresolved.
If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.