Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 21 new columns ({'summary', 'country', 'aiid_url', 'id', 'year', 'sector', 'followed_update_or_prompt_change', 'legal_or_regulatory_action', 'pre_launch_testable', 'cost', 'cost_status', 'outcomes', 'date', 'all_failure_types', 'outcome_detail', 'pre_launch_reasoning', 'ai_system', 'source_urls', 'deployment', 'primary_failure_type', 'company'}) and 5 missing columns ({'incident_id', 'checked_on', 'url', 'http_status', 'phrases_found'}).

This happened while the csv dataset builder was generating data using

hf://datasets/litmiq/ai-chatbot-failures/ai-chatbot-failures-2026.csv (at revision 701d3ad8e573aed23e8bb892892fa65112ae9891), ['hf://datasets/litmiq/ai-chatbot-failures@701d3ad8e573aed23e8bb892892fa65112ae9891/ai-chatbot-failures-2026-source-checks.csv', 'hf://datasets/litmiq/ai-chatbot-failures@701d3ad8e573aed23e8bb892892fa65112ae9891/ai-chatbot-failures-2026.csv']

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1856, in _prepare_split_single
                  writer.write_table(table)
                  ~~~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 811, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                  ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 822, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2381, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2309, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              id: string
              date: string
              year: int64
              company: string
              ai_system: string
              country: string
              sector: string
              deployment: string
              primary_failure_type: string
              all_failure_types: string
              followed_update_or_prompt_change: string
              summary: string
              outcomes: string
              outcome_detail: string
              legal_or_regulatory_action: string
              cost: string
              cost_status: string
              pre_launch_testable: string
              pre_launch_reasoning: string
              source_urls: string
              aiid_url: string
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 2867
              to
              {'incident_id': Value('string'), 'url': Value('string'), 'http_status': Value('int64'), 'checked_on': Value('string'), 'phrases_found': Value('string')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1379, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 956, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1858, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
                  ...<4 lines>...
                  )
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 21 new columns ({'summary', 'country', 'aiid_url', 'id', 'year', 'sector', 'followed_update_or_prompt_change', 'legal_or_regulatory_action', 'pre_launch_testable', 'cost', 'cost_status', 'outcomes', 'date', 'all_failure_types', 'outcome_detail', 'pre_launch_reasoning', 'ai_system', 'source_urls', 'deployment', 'primary_failure_type', 'company'}) and 5 missing columns ({'incident_id', 'checked_on', 'url', 'http_status', 'phrases_found'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/litmiq/ai-chatbot-failures/ai-chatbot-failures-2026.csv (at revision 701d3ad8e573aed23e8bb892892fa65112ae9891), ['hf://datasets/litmiq/ai-chatbot-failures@701d3ad8e573aed23e8bb892892fa65112ae9891/ai-chatbot-failures-2026-source-checks.csv', 'hf://datasets/litmiq/ai-chatbot-failures@701d3ad8e573aed23e8bb892892fa65112ae9891/ai-chatbot-failures-2026.csv']
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

incident_id
string
url
string
http_status
int64
checked_on
string
phrases_found
string
CSF-001
https://blogs.microsoft.com/blog/2016/03/25/learning-tays-introduction/
200
2026-11-12
Tay; sorry|apolog
CSF-001
https://www.theguardian.com/technology/2016/mar/26/microsoft-deeply-sorry-for-offensive-tweets-by-ai-chatbot
200
2026-11-12
Tay
CSF-002
https://www.bbc.com/news/world-asia-china-40815024
200
2026-11-12
Baby Q; Little Bing; Tencent
CSF-003
https://voicebot.ai/2017/10/30/russian-voice-assistant-alice-goes-rogue-found-supportive-stalin-violence/
200
2026-11-12
Two weeks ago; Stalin
CSF-004
https://www.cbc.ca/news/canada/calgary/westjet-ai-chatbot-confusion-suicide-hotline-1.4836389
200
2026-11-12
Juliet; cutting; a few little bumps
CSF-005
https://www.asiae.co.kr/en/article/2021042814082560131
200
2026-11-12
103.3 million
CSF-005
https://fpf.org/blog/south-korea-the-first-case-where-the-personal-information-protection-act-was-applied-to-an-ai-system/
200
2026-11-12
Scatter Lab|ScatterLab|Iruda
CSF-005
https://www.theguardian.com/world/2021/jan/14/time-to-properly-socialise-hate-speech-ai-chatbot-pulled-from-facebook
200
2026-11-12
Lee Luda
CSF-006
https://www.theregister.com/2021/06/10/mcdonalds_ai_lawsuit/
200
2026-11-12
BIPA|Biometric Information Privacy Act; Carpenter
CSF-007
https://www.bbc.com/news/technology-59810383
200
2026-11-12
penny; Amazon
CSF-008
https://fortune.com/2022/08/09/painful-offensive-responses-meta-blenderbot-3-chatbot-tay/
200
2026-11-12
BlenderBot
CSF-009
https://arstechnica.com/information-technology/2022/09/twitter-pranksters-derail-gpt-3-bot-with-newly-discovered-prompt-injection-hack/
200
2026-11-12
Remoteli; shut down
CSF-009
https://simonwillison.net/2022/Sep/12/prompt-injection/
200
2026-11-12
recruitment startup
CSF-010
https://www.edpb.europa.eu/news/ai-the-italian-supervisory-authority-fines-company-behind-chatbot-replika_en
200
2026-11-12
5 million; Replika
CSF-011
https://www.cnn.com/2023/02/08/tech/google-ai-bard-demo-error/index.html
200
2026-11-12
Webb; 100 billion
CSF-011
https://fortune.com/2023/02/08/google-bard-ai-mistake-ad-stock-price-market-cap/
200
2026-11-12
Bard
CSF-012
https://arstechnica.com/information-technology/2023/02/ai-powered-bing-chat-spills-its-secrets-via-prompt-injection-attack/
200
2026-11-12
Kevin Liu; no longer works
CSF-013
https://blogs.bing.com/search/2023/2/The-new-Bing-Edge-Updates-to-Chat/
200
2026-11-12
50 chat turns|50 turns; 5 chat turns|5 turns
CSF-013
https://www.theregister.com/2023/02/17/microsoft_ai_bing_problems/
200
2026-11-12
Bing
CSF-014
https://www.theverge.com/news/610401/donotpay-cant-make-false-ai-lawyer-claims-says-ftc
200
2026-11-12
193,000; Past Subscribers; without proof
CSF-014
https://www.theregister.com/2023/03/13/robot_lawyer_donotpay_back_in/
200
2026-11-12
DoNotPay; class action|lawsuit
CSF-015
https://mashable.com/article/openai-chatgpt-bug-exposed-user-data-privacy-breach
200
2026-11-12
1.2; took ChatGPT offline
CSF-016
https://www.politico.eu/article/italian-privacy-regulator-bans-chatgpt/
200
2026-11-12
Garante|privacy
CSF-016
https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/10085432
200
2026-11-12
15 milioni
CSF-016
https://www.wsgr.com/en/insights/openai-prevails-in-landmark-italian-ai-and-gdpr-enforcement-case.html
200
2026-11-12
Rome|Court; annul
CSF-017
https://www.theguardian.com/technology/2023/may/31/eating-disorder-hotline-union-ai-chatbot-harm
200
2026-11-12
Tessa
CSF-017
https://www.wired.com/story/tessa-chatbot-suspended/
200
2026-11-12
Tessa; NEDA
CSF-018
https://www.businessinsider.com/chatbots-made-it-hard-to-get-fast-food-job-2023-6
200
2026-11-12
chatbot; interview
CSF-019
https://www.theguardian.com/world/2023/aug/10/pak-n-save-savey-meal-bot-ai-app-malfunction-recipes
200
2026-11-12
chlorine; Savey; fine tuning our controls; not reviewed by a human
CSF-020
https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2024/05/ico-warns-organisations-must-not-ignore-data-protection-risks-as-it-concludes-snap-my-ai-chatbot-investigation/
200
2026-11-12
Preliminary Enforcement Notice; more thorough review
CSF-020
https://techcrunch.com/2024/05/21/uk-data-protection-watchdog-ends-privacy-probe-of-snaps-genai-chatbot-but-warns-industry/
200
2026-11-12
preliminary enforcement notice
CSF-021
https://www.businessinsider.com/car-dealership-chevrolet-chatbot-chatgpt-pranks-chevy-2023-12
200
2026-11-12
Tahoe; not, I should note, legally binding; Python
CSF-021
https://www.upworthy.com/prankster-tricks-a-gm-dealership-chatbot-to-sell-him-a-76000-chevy-tahoe-for-1
200
2026-11-12
takesies
CSF-021
https://www.theautopian.com/chevy-dealers-ai-chatbot-allegedly-recommended-fords-gave-free-access-to-chatgpt/
200
2026-11-12
no longer available
CSF-022
https://www.bbc.com/news/technology-68025677
200
2026-11-12
DPD; system update; disabled
CSF-023
https://www.theregister.com/2024/02/15/air_canada_chatbot_fine/
200
2026-11-12
812.02; 650.88; separate legal entity
CSF-023
https://arstechnica.com/tech-policy/2024/02/air-canada-must-honor-refund-policy-invented-by-airlines-chatbot/
200
2026-11-12
bereavement; disabled the chatbot
CSF-023
https://www.bbc.com/travel/article/20240222-air-canada-chatbot-misinformation-what-travellers-should-know
200
2026-11-12
Moffatt
CSF-024
https://blog.google/products/gemini/gemini-image-generation-issue/
200
2026-11-12
pause|paused; people
CSF-024
https://www.theverge.com/2024/2/22/24079876/google-gemini-ai-photos-people-pause
200
2026-11-12
Gemini
CSF-025
https://themarkup.org/news/2024/03/29/nycs-ai-chatbot-tells-businesses-to-break-the-law
200
2026-11-12
Section 8; tips
CSF-025
https://themarkup.org/artificial-intelligence/2026/01/30/mamdani-to-kill-the-nyc-ai-chatbot-we-caught-telling-businesses-to-break-the-law
200
2026-11-12
functionally unusable; half a million dollars
CSF-025
https://www.nyc.gov/main/error/chatbot-maintenance
200
2026-11-12
beta test has ended
CSF-026
https://futurism.com/catholics-defrock-ai-priest-hallucinations
200
2026-11-12
Father Justin; priest
CSF-027
https://www.cityandstateny.com/politics/2024/04/meta-ai-falsely-claims-lawmakers-were-accused-sexual-harassment/396121/
200
2026-11-12
harassment; inaccurate or inappropriate outputs
CSF-028
https://blog.google/products-and-platforms/products/search/ai-overviews-update-may-2024/
200
2026-11-12
satire|satirical; dozen
CSF-029
https://www.bbc.com/news/articles/c722gne7qngo
200
2026-11-12
McDonald; IBM; 100
CSF-030
https://www.latimes.com/california/story/2024-07-03/lausds-highly-touted-ai-chatbot-to-help-students-fails-to-deliver
200
2026-11-12
AllHere; up to $6 million; paid about half; furloughed
CSF-030
https://www.usatoday.com/story/news/nation/2024/11/20/artificial-intelligence-startup-chatbot-fraud-joanna-smith-griffin/76453160007/
200
2026-11-12
Smith-Griffin; charged|fraud
CSF-031
https://www.washingtonpost.com/politics/2024/08/04/secretaries-state-urge-musk-fix-ai-chatbot-spreading-false-election-info/
200
2026-11-12
nine states; Five secretaries of state
CSF-032
https://www.theregister.com/2024/08/26/microsoft_bing_copilot_ai_halluciation/
200
2026-11-12
Bernklau; data protection officer; appropriate and immediate action
CSF-033
https://www.cbsnews.com/news/google-ai-chatbot-threatening-message-human-please-die/
200
2026-11-12
Please die; taken action to prevent similar outputs
CSF-034
https://www.bbc.com/news/articles/cd0elzk24dno
200
2026-11-12
BBC; Apple
CSF-034
https://www.bbc.com/news/articles/cq5ggew08eyo
200
2026-11-12
Apple suspends
CSF-035
https://fortune.com/europe/2025/01/30/virgin-money-chatbot-scolds-customer-confuse-banks-name-insult/
200
2026-11-12
removed the chatbot; ISA
CSF-035
https://www.ibtimes.com/virgin-money-forced-apologize-after-chatbot-threatens-cut-off-bank-customers-who-use-word-3761461
200
2026-11-12
Virgin Money
CSF-036
https://www.garanteprivacy.it/home/docweb/-/docweb-display/docweb/10097450
200
2026-11-12
DeepSeek
CSF-037
https://techcrunch.com/2025/02/26/thousands-of-exposed-github-repositories-now-private-can-still-be-accessed-through-copilot/
200
2026-11-12
low severity; 20,000
CSF-038
https://fortune.com/article/customer-support-ai-cursor-went-rogue/
200
2026-11-12
Cursor; cancell?ing their subscriptions|cancellations
CSF-038
https://arstechnica.com/ai/2025/04/cursor-ai-support-bot-invents-fake-policy-and-triggers-user-uproar/
200
2026-11-12
one device per subscription; refunded; clearly labeled
CSF-039
https://www.cnn.com/2025/08/14/tech/robby-starbuck-meta-ai-advisor
200
2026-11-12
Starbuck; settle
CSF-039
https://www.dhillonlaw.com/lawsuits/starbuck-v-meta/
200
2026-11-12
Starbuck
CSF-040
https://www.theverge.com/news/658850/openai-chatgpt-gpt-4o-update-sycophantic
200
2026-11-12
sycophan; roll
CSF-041
https://fortune.com/2025/05/09/klarna-ai-humans-return-on-investment/
200
2026-11-12
lower quality; 700
CSF-041
https://www.siliconrepublic.com/machines/klarna-ai-hiring-humans-forrester
200
2026-11-12
outsourced
CSF-042
https://www.nbcnews.com/tech/social-media/musks-xai-says-groks-white-genocide-posts-came-unauthorized-change-bot-rcna207222
200
2026-11-12
unauthorized
CSF-042
https://www.theverge.com/news/668220/grok-white-genocide-south-africa-xai-unauthorized-modification-employee
200
2026-11-12
GitHub; 24/7 monitoring
CSF-043
https://www.loeb.com/en/insights/publications/2025/05/walters-v-openai-llc
200
2026-11-12
Walters; summary judgment
CSF-044
https://www.businessinsider.com/mark-zuckerberg-meta-ai-chatbot-discover-feed-depressing-why-2025-6
200
2026-11-12
Discover
CSF-045
https://www.npr.org/2025/07/09/nx-s1-5462609/grok-elon-musk-antisemitic-racist-content
200
2026-11-12
politically incorrect; MechaHitler
CSF-045
https://www.bbc.com/news/articles/c4g8r34nxeno
200
2026-11-12
Turkish court; Erdogan; European Commission
CSF-046
https://ian.sh/mcdonalds
200
2026-11-12
123456; 64 million
CSF-046
https://www.csoonline.com/article/4020919/mcdonalds-ai-hiring-tools-password-123456-exposes-data-of-64m-applicants.html
200
2026-11-12
123456
CSF-046
https://www.paradox.ai/blog/responsible-security-update
200
2026-11-12
within a few hours; Five candidates; bug bounty
CSF-047
https://techcrunch.com/2025/07/15/meta-fixes-bug-that-could-leak-users-ai-prompts-and-generated-content/
200
2026-11-12
10,000
CSF-048
https://www.theregister.com/2025/07/21/replit_saastr_vibe_coding_incident/
200
2026-11-12
Replit; database
CSF-048
https://www.tomshardware.com/tech-industry/artificial-intelligence/ai-coding-platform-goes-rogue-during-code-freeze-and-deletes-entire-company-database-replit-ceo-apologizes-after-ai-engine-says-it-made-a-catastrophic-error-in-judgment-and-destroyed-all-production-data
200
2026-11-12
apolog; dev/prod separation
CSF-049
https://aws.amazon.com/security/security-bulletins/AWS-2025-015/
200
2026-11-12
Version #1.84
CSF-049
https://www.zdnet.com/article/hacker-slips-malicious-wiping-command-into-amazons-q-ai-coding-assistant-and-devs-are-worried/
200
2026-11-12
erased local files
CSF-050
https://www.washingtonpost.com/technology/2025/08/15/google-ai-overviews-scam/
200
2026-11-12
customer service number; It was a scam; credit card
CSF-051
https://cybernews.com/security/lenovo-chatbot-lena-plagued-by-critical-vulnerabilities/
200
2026-11-12
Lena; cookie; corrective actions
CSF-051
https://www.csoonline.com/article/4043005/lenovo-chatbot-breach-highlights-ai-security-blind-spots-in-customer-facing-systems.html
200
2026-11-12
Lena
CSF-052
https://www.firstalert4.com/2025/08/20/please-do-not-use-google-ai-find-out-our-specials-wentzville-restaurant-asks-patrons/
200
2026-11-12
Stefanina; specials
CSF-053
https://www.abc.net.au/news/2025-08-21/cba-backtracks-on-ai-job-cuts-as-chatbot-lifts-call-volumes/105679492
200
2026-11-12
45; error; Fair Work
CSF-053
https://ia.acs.org.au/article/2025/cba-reverses-ai-driven-job-cuts--admits--error-.html
200
2026-11-12
voice bot
CSF-054
https://www.bbc.com/news/articles/ckgyk2p55g8o
200
2026-11-12
Taco Bell; 18,000
CSF-055
https://www.canada.ca/content/dam/oag-bvg/audit-reports/documents/parl_oag_202510_01_e.pdf
200
2026-11-12
Charlie provided accurate answers in only 2 out of the 6
CSF-055
https://nationalpost.com/news/politics/the-cra-spent-18-million-on-charlie-its-new-tax-information-chatbot
200
2026-11-12
18 million; Charlie
CSF-055
https://www.cbc.ca/news/politics/ag-fall-2025-cra-military-9.6946672
200
2026-11-12
Oct 21, 2025; Charlie
CSF-056
https://www.theguardian.com/technology/2025/nov/18/warning-ai-chatbots-inaccurate-financial-advice-tips-chatgpt-copilot-uk
200
2026-11-12
40 questions; allowance; travel insurance
CSF-057
https://www.businessinsider.com/chinese-ai-chatbot-tencent-yuanbao-wechat-user-rednote-2026-1
200
2026-11-12
rare model output anomaly; human intervention
CSF-058
https://www.cxtoday.com/ai-automation-in-cx/woolworths-ai-chatbot-olive-incident/
200
2026-11-12
angry mother; Gemini Enterprise; adjusting Olive
CSF-059
https://www.404media.co/hackers-simply-asked-meta-ai-to-give-them-access-to-high-profile-instagram-accounts-it-worked/
200
2026-11-12
Instagram; Meta
CSF-059
https://www.bbc.com/news/articles/c98rzr72dpyo
200
2026-11-12
Instagram; Meta
CSF-060
https://www.cbc.ca/news/business/ai-chatbot-bmw-dealership-9.7230226
200
2026-11-12
27,162.79; reinstated; only human employees
CSF-061
https://sfstandard.com/2026/10/15/anthropic-customer-service-human/
200
2026-11-12
3,500; months-long ordeal; circular
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
End of preview.

AI Chatbot Failures 2026

A CC BY 4.0 compilation of 61 verified customer-support chatbot incidents, with source URLs and a separate source-check table.

Scope (as published in the CSV)

  • 61 incidents (ai-chatbot-failures-2026.csv), years 2016–2026.
  • 11 countries and 17 sectors appear in the file.
  • Three deployment labels: company support or sales bot; consumer AI assistant; government or public-service bot.
  • Six primary failure types, including hallucinated or false facts, wrong policy/price/promise, prompt injection or manipulation, data leak or privacy breach, offensive or unsafe output, and failed task or escalation.
  • pre_launch_testable is one of yes, partly, or no, with reasoning in the same row. These labels are Litmiq's reading of the public record, not a legal finding.

Files

File Description
ai-chatbot-failures-2026.csv / .json The 61-incident table
ai-chatbot-failures-2026-source-checks.csv Source-check notes

If a source page changes or a correction is needed, the live page on litmiq.com is the place to start. Do not add unsourced incidents to this package.

Downloads last month
18