{"data":{"slug":"kennethleungty-failed-ml","name":"Failed-ML","tagline":"Compilation of high-profile real-world examples of failed machine learning projects","github_url":"https://github.com/kennethleungty/Failed-ML","owner":"kennethleungty","repo":"Failed-ML","owner_avatar_url":"https://avatars.githubusercontent.com/u/16691614?v=4","primary_language":null,"stars":751,"forks":52,"topics":["ai","artificial-intelligence","classification","computer-vision","data-engineering","data-quality","data-science","deep-learning","failed-data-science","failed-machine-learning","failed-ml","fml","forecasting","machine-learning","ml","natural-language-processing","production","recsys","regression"],"archived":false,"github_pushed_at":"2024-06-14T08:32:21+00:00","maintenance_label":"Dormant","stars_delta_30d":-2,"url":"https://www.graphcanon.com/tools/kennethleungty-failed-ml","markdown_url":"https://www.graphcanon.com/tools/kennethleungty-failed-ml.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/kennethleungty-failed-ml","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=kennethleungty-failed-ml","description":"Compilation of high-profile real-world examples of failed machine learning projects","homepage_url":"https://towardsdatascience.com/when-ai-goes-astray-high-profile-machine-learning-mishaps-in-the-real-world-26bd58692195","license":"MIT","open_issues":0,"watchers":17,"ai_summary":"Provides insights into the failures of various ML projects across different domains to aid in understanding common pitfalls and mistakes in implementing machine learning solutions.","readme_excerpt":"# Failed Machine Learning (FML)\n## High-profile real-world examples of failed machine learning projects\n
\n\n

\n \n

\n\n\n\n\n> ### “Success is not final, failure is not fatal. It is the courage to continue that counts.” - Winston Churchill\n
\n\nIf you are looking for examples of how ML can fail despite all its incredible potential, you have come to the right place. Beyond the wonderful success stories of applied machine learning, here is a list of failed projects which we can learn a lot from. \n

\n\n\n
\n\n## Contents\n1. [Classic Machine Learning](#classical-machine-learning)\n2. [Computer Vision](#computer-vision)\n3. [Forecasting](#forecasting)\n4. [Image Generation](#image-generation)\n5. [Natural Language Processing](#natural-language-processing)\n6. [Recommendation Systems](#recommendation-systems)\n\n\n___\n\n## Classic Machine Learning\n| Title | Description |\n| --- | --- |\n| [Amazon AI Recruitment System](https://finance.yahoo.com/news/amazon-reportedly-killed-ai-recruitment-100042269.html) | AI-powered automated recruitment system canceled after evidence of discrimination against female candidates |\n| [Genderify - Gender identification tool](https://syncedreview.com/2020/07/30/ai-powered-genderify-platform-shut-down-after-bias-based-backlash/) | AI-powered tool designed to identify gender based on fields like name and email address was shut down due to built-in biases and inaccuracies | \n| [Leakage and the Reproducibility Crisis in ML-based Science](https://arxiv.org/pdf/2207.07048.pdf) | A team at Princeton University found 20 reviews across 17 scientific fields that discovered significant errors (e.g., data leakage, no train-test split) in 329 papers that use ML-based science |\n| [COVID-19 Diagnosis and Triage Models](https://www.technologyreview.com/2021/07/30/1030329/machine-learning-ai-failed-covid-hospital-diagnosis-pandemic/) | Hundreds of predictive models were developed to diagnose or triage COVID-19 patients faster, but ultimately none of them were fit for clinical use, and some were potentially harmful | \n| [COMPAS Recidivism Algorithm](https://www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm) | Florida’s recidivism risk system found evidence of racial bias | \n| [Pennsylvania Child Welfare Screening Tool](https://apnews.com/article/child-welfare-algorithm-investigation-9497ee937e0053ad4144a86c68241ef1) | The predictive algorithm (which helps identify which families are to be investigated by social workers for child abuse and neglect) flagged a disproportionate number of Black children for 'mandatory' neglect investigations. | \n| [Oregon Child Welfare Screening Tool](https://apnews.com/article/politics-technology-pennsylvania-child-abuse-1ea160dc5c2c203fdab456e3c2d97930) | A similar predictive tool to the one in Pennsylvania, the AI algorithm for child welfare in Oregon was also stopped a month after the Pennsylvania report | \n| [U.S. Healthcare System Health Risk Prediction](https://www.science.org/doi/full/10.1126/science.aax2342) | A widely used algorithm to predict healthcare needs exhibited racial bias where for a given risk score, black patients are considerably sicker than white patients | \n| [Apple Card Credit Card](https://www.theverge.com/2019/11/11/20958953/apple-credit-card-gender-discrimination-algorithms-black-box-investigation) | Apple’s new credit card (created in partnership with Goldman Sachs) is being investigated by financial regulators after customers complained that the card’s lending algorithms discriminated against women, where the credit line offered by a male customer's Apple Card was 20 times higher than that offered to his spouse | \n\n___\n\n## Computer Vision\n| Title | Description |\n| --- | --- |\n| [Inverness Automated Football Camera System](https://www.theverge.com/tldr/2020/11/3/21547392","github_created_at":"2022-08-15T14:46:37+00:00","created_at":"2026-07-11T12:27:46.507799+00:00","updated_at":"2026-09-23T06:00:25.781985+00:00","categories":[{"slug":"evaluation-observability","name":"Evaluation & Observability","url":"https://www.graphcanon.com/categories/evaluation-observability","markdown_url":"https://www.graphcanon.com/categories/evaluation-observability.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/evaluation-observability"}],"tags":[{"slug":"ai","name":"ai"},{"slug":"artificial-intelligence","name":"artificial-intelligence"},{"slug":"classification","name":"classification"},{"slug":"computer-vision","name":"computer-vision"},{"slug":"data-engineering","name":"data-engineering"},{"slug":"data-quality","name":"data-quality"},{"slug":"data-science","name":"data-science"},{"slug":"deep-learning","name":"deep-learning"}],"trust":{"provenance":{"is_fork":false,"github_id":525025357,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-23T06:00:24.829Z","maintenance":{"label":"Dormant","score":18,"methodology":"github_public_v1","releases_90d":0,"days_since_push":830,"last_release_at":null,"stars_delta_30d":-2,"open_issues_delta_30d":0},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T12:27:48.241Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-09-23T06:00:25.495Z"},"license_spdx":{"value":"MIT","source":"github.license","observed_at":"2026-09-23T06:00:25.495Z"}},"decision_facts":{"hosting":{"model":"self_hosted"},"pricing":{"model":"freemium","summary":"Open source under MIT license but no additional paid features are mentioned."},"requirements":{"notes":["Not a software tool that requires installation. Informational repository intended for reading and learning."],"min_ram_gb":null,"requires_docker":false},"constraints":{"min_ram_gb":null,"hosting_model":"self_hosted","pricing_model":"freemium","requires_docker":false},"when_to_use":["When you seek specific historical examples of where machine learning application went wrong, aiding in understanding the potential mistakes and challenges one might face.","For educating teams or stakeholders on why certain ML projects failed, providing real-world context to abstract concepts such as data bias, model validation issues, and ethical concerns."],"when_not_to_use":["If you need a prescriptive guide for solving your current project's technical problems rather than learning from industry-wide mistakes.","When looking for detailed quantitative metrics on failed projects, Failed-ML focuses more on high-level analysis of failure causes rather than specific performance numbers."],"source":"enrich:decision_facts","observed_at":"2026-07-16T20:34:18.512Z"},"constraint_facets":{"min_ram_gb":null,"hosting_model":"self_hosted","pricing_model":"freemium","requires_docker":false},"decision_summary":[{"label":"Hosting","value":"self hosted"},{"label":"Pricing","value":"freemium - Open source under MIT license but no additional paid features are mentioned."},{"label":"Requirements","value":"Not a software tool that requires installation. Informational repository intended for reading and learning."},{"label":"Adopt for","value":"Failed-ML is compiled around high-profile ML project failures across domains and includes detailed insights from these cases to help understand common pitfalls in implementing machine learning systems."}]}}