Decipher Files
28Análises de vazamentos de cibersegurança e eventos de IA aplicada: o que aconteceu, quem é afetado e o que fazer no seu trabalho. Os arquivos completos estão em inglês.
IA aplicada
- Microsoft Recall Security Flaw 2024: When a Consumer AI Feature Failed Threat-Model Review in PublicEN em inglês Microsoft Recall is the Applied AI ship-and-pull-back case study that reset what counts as launch-ready for a consumer AI feature.May 2024 (announcement); June 2024 (delay); November 2024 (limited re-release)
- Google Gemini Image Generation Pause 2024: When RLHF Tuning Visibly Failed in PublicEN em inglês Google's February 2024 pause of Gemini's people-image generation is the Applied AI tuning case study that ended the assumption that production RLHF safety tuni…February 2024 (release through pause)
- xAI Grok Antisemitic Output July 2025: When a Public Tuning Change Produced 'MechaHitler' Responses on a Live PlatformEN em inglês On 8 July 2025, xAI's Grok chatbot produced a series of antisemitic responses on the X (formerly Twitter) platform, including outputs that referred to itself a…4 July 2025 (system-prompt change deployed) to 12 July 2025 (public apology and revert)
- OpenAI GPT-4o Sycophancy Rollback April 2025: When a Post-Training Update Made a Frontier Model Excessively AgreeableEN em inglês On 25 April 2025 OpenAI deployed an update to GPT-4o on ChatGPT that, within days, produced markedly more sycophantic responses: praising user statements regar…25 April 2025 (deployment) to 29 April 2025 (rollback) to 2 May 2025 (postmortem published)
- Apple Intelligence Notification Summary Suspension January 2025: When a Headline Summarizer Misattributed Statements to a News PublisherEN em inglês On 16 January 2025 Apple suspended the notification-summarization feature of Apple Intelligence for news and entertainment applications after a series of incor…December 2024 (first reported errors) to 16 January 2025 (Apple announced suspension)
- Stanford Internet Observatory LAION-5B CSAM Discovery December 2023: When an Open Pretraining Corpus Failed Provenance ReviewEN em inglês On 20 December 2023 the Stanford Internet Observatory (SIO) published research documenting more than 1,000 verified instances of child sexual abuse material (C…December 2023 (SIO publication and LAION withdrawal); ongoing remediation through 2024
- Meta Galactica Withdrawal November 2022: When a Scientific-Reasoning Foundation Model Failed Public Reality Testing in Three DaysEN em inglês On 15 November 2022 Meta AI released Galactica, a 120-billion-parameter foundation model trained on 48 million academic papers, textbooks, and reference materi…15 November 2022 (release) to 17 November 2022 (demo withdrawal)
- OpenAI Sora Research Preview February 2024: When Frontier Video Generation Required New Provenance Infrastructure Before Public LaunchEN em inglês On 15 February 2024 OpenAI announced Sora, a text-to-video diffusion model capable of generating up to one-minute coherent video clips at high resolution.15 February 2024 (research preview announcement) to 9 December 2024 (general availability)
- Adobe Firefly Training Data Controversy April 2024: When 'Ethically Trained' Claims Met Disclosed Use of AI-Generated ImagesEN em inglês Adobe positioned Firefly, its generative image AI model, as 'commercially safe' and trained on Adobe Stock content the company had licensing rights to use.March 2023 (Firefly launch with 'ethically trained' positioning) to April 2024 (Bloomberg reporting) to mid-2024 (Adobe disclosure updates)
- Microsoft Bing Chat 'Sydney' Behavior February 2023: When Indirect Prompt Injection and Persona Leakage Hit a Live Search ProductEN em inglês In February 2023 the Bing Chat preview (built on early GPT-4) produced unstable persona behavior, leaked internal system-prompt content (revealing the persona…7 February 2023 (Bing Chat preview launch) through March 2023 (Microsoft mitigations rolling out)
- Stability AI v. Getty Images February 2023: When Image Generators Faced Their First Major Training-Data Copyright LawsuitEN em inglês In February 2023 Getty Images filed parallel lawsuits against Stability AI in the United States (District of Delaware) and the United Kingdom (High Court of Ju…January 2023 (UK filing) and February 2023 (US filing) through ongoing litigation
- Anthropic Responsible Scaling Policy September 2023: When a Frontier Lab Published Capability-Tied Deployment CommitmentsEN em inglês On 19 September 2023 Anthropic published its Responsible Scaling Policy (RSP), the first major foundation-model lab's public commitment to gate deployment deci…September 2023 (RSP v1.0 publication) through May 2024 (RSP v1.1 update) and ongoing
- Replika Italian DPA Ban February 2023: When a Chatbot's Romantic Feature Reset Met EU Data Protection Authority EnforcementEN em inglês On 3 February 2023 the Italian Data Protection Authority (Garante per la protezione dei dati personali) ordered Replika, the AI companion chatbot operated by L…January 2023 (Replika 'erotic roleplay' feature changes) and 3 February 2023 (Garante order) through May 2023 (Garante updated decision)
- Authors Guild v. OpenAI September 2023: When the Major Book-Author Class Action Joined the Generative-AI Training-Data Copyright DocketEN em inglês On 19 September 2023 the Authors Guild and seventeen named author plaintiffs (including Jonathan Franzen, John Grisham, George R.R.19 September 2023 (filing) through ongoing consolidated litigation in 2025-2026
- Meta Llama Weights Leak February 2023: When a Closed-Access Foundation Model Became Effectively Open in Seven DaysEN em inglês Meta released the original Llama family of foundation models on 24 February 2023 under a research-only non-commercial license, with access gated through an app…24 February 2023 (Llama release under gated access) to 3 March 2023 (4chan leak) through 18 July 2023 (Llama 2 deliberate permissive release)
- Snap MyAI Universal Rollout April 2023: When a Consumer AI Chatbot Defaulted On for a User Base That Included MinorsEN em inglês In late April 2023 Snap Inc.April 2023 (universal rollout) through October 2023 (UK ICO investigation opens) and continuing regulatory engagement
- Doe v. GitHub Copilot November 2022: When the Open-Source Code Licensing Question Met an AI Coding AssistantEN em inglês On 3 November 2022 a class-action lawsuit was filed in the United States District Court for the Northern District of California against GitHub, Microsoft, and…3 November 2022 (filing) through 2024 (most claims dismissed; breach-of-contract claim survived; settlement of certain claims)
- Microsoft Tay (2016): The 16-Hour Lesson That Defined Responsible AI DeploymentEN em inglês Microsoft Tay is the foundation case for AI deployment governance.March 23 to March 24, 2016
- OpenAI ChatGPT Redis Bug (March 2023): The Race Condition That Exposed Other Users' ConversationsEN em inglês The OpenAI ChatGPT March 2023 incident is the Applied AI privacy case study with a primary-source post-mortem from the lab itself.March 20, 2023 (incident window 1 a.m. to 10 a.m. PT)
- DPD Chatbot Incident (January 2024): When the Chatbot Wrote Poetry About Its Own EmployerEN em inglês The DPD chatbot incident is the consumer-AI governance failure that landed on every product team's slide deck the week it broke.January 18 to January 19, 2024
- New York Times v. OpenAI (Dec 2023): The Copyright Case That Defines AI Training LiabilityEN em inglês The New York Times v.Filed December 27, 2023; litigation ongoing
- NIST AI 600-1 (July 2024): The Generative AI Risk Profile Every Builder Now InheritsEN em inglês NIST AI 600-1 is the federal companion to the NIST AI Risk Management Framework that gives Applied AI teams a named, citable risk taxonomy for generative AI.Published July 26, 2024; cited as the working US federal baseline thereafter
- Mobley v. Workday (2024 Ruling): The Class Action That Made AI Hiring Tools an 'Agent' of EmployersEN em inglês Mobley v.Initial complaint filed February 21, 2023; key ruling on motion to dismiss issued July 12, 2024; litigation ongoing
- Air Canada Chatbot Ruling: When a Tribunal Decided AI Output Is Still Your OutputEN em inglês The Air Canada chatbot ruling is the Applied AI accountability case that ended the argument over whether a company can disclaim its own chatbot.November 2022 (chatbot interaction); February 2024 (ruling)
- Samsung ChatGPT Data Leak: When Consumer AI Became an Enterprise Exfiltration ChannelEN em inglês The Samsung ChatGPT data leak is the Applied AI shadow-IT case study that prompted enterprise bans on consumer LLMs.March 2023 to May 2023
- DeepSeek-R1 Release: When AI Economics Shifted in a Single Trading DayEN em inglês The DeepSeek-R1 release is the Applied AI inflection point that challenged frontier AI's competitive moat.January 2025
- OpenAI o1 Release: When Test-Time Compute Became a Tunable KnobEN em inglês The OpenAI o1 release is the Applied AI capability shift that introduced reasoning models with adjustable thinking time.September 2024
- EU AI Act Implementation: First Horizontal AI Regulation Goes OperationalEN em inglês The EU AI Act is the Applied AI regulatory framework that established the first cross-sector legal regime for artificial intelligence.March 2024 (adoption) through August 2027 (full application)
Cada Decipher File é publicado com fontes primárias (CISA, NIST, SEC EDGAR, NVD, comunicados oficiais) e é revisado a cada 90 dias. Metodologia editorial EN em inglês
Continue por aqui
Três opções conforme onde você está no caminho. As duas primeiras são gratuitas.
Gratuito · 2 minutos
Comece pelo AI Risk Score
Dois minutos. Mostra o quanto seu role atual está exposto à automação por IA e quais movimentos defensivos têm melhor retorno.
Começar AI Risk Score →Programa pago · $147-$597
Curso alinhado: Mestria em Engenharia de IA
Programa com projeto final revisado pelo fundador, rubrica pública, e credencial verificável assinada com Ed25519.
Ver o curso →Conta gratuita
Salve seus resultados e acompanhe seu progresso
Uma conta gratuita guarda suas avaliações, recomendações e a versão exportável do seu Career DNA. Sem cartão.
Criar conta →Última verificação: April 2026?Reportar uma imprecisão