Skip to Content
logologo
AI Incident Database
Open TwitterOpen RSS FeedOpen FacebookOpen LinkedInOpen GitHub
Open Menu
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse

Incident 72: Facebook translates 'good morning' into 'attack them', leading to arrest

Description: Facebook's automatic language translation software incorrectly translated an Arabic post saying "Good morning" into Hebrew saying "hurt them," leading to the arrest of a Palestinian man in Beitar Illit, Israel.

Tools

New ReportNew ReportNew ResponseNew ResponseDiscoverDiscoverView HistoryView History

Entities

View all entities
Alleged: Facebook developed and deployed an AI system, which harmed unnamed Palestinian Facebook user , Palestinian Facebook users , Arabic-speaking Facebook users and Facebook users.

Incident Stats

Incident ID
72
Report Count
26
Incident Date
2017-10-17
Editors
Sean McGregor, Khoa Lam
Applied Taxonomies
CSETv0, GMF, CSETv1, MIT

CSETv1 Taxonomy Classifications

Taxonomy Details

Incident Number

The number of the incident in the AI Incident Database.
 

72

AI Tangible Harm Level Notes

Notes about the AI tangible harm level assessment
 

A Palestinian man was arrested and questioned for several hours due to translation error on Facebook.

Special Interest Intangible Harm

An assessment of whether a special interest intangible harm occurred. This assessment does not consider the context of the intangible harm, if an AI was involved, or if there is characterizable class or subgroup of harmed entities. It is also not assessing if an intangible harm occurred. It is only asking if a special interest intangible harm occurred.
 

yes

Date of Incident Year

The year in which the incident occurred. If there are multiple harms or occurrences of the incident, list the earliest. If a precise date is unavailable, but the available sources provide a basis for estimating the year, estimate. Otherwise, leave blank. Enter in the format of YYYY
 

2017

Date of Incident Month

The month in which the incident occurred. If there are multiple harms or occurrences of the incident, list the earliest. If a precise date is unavailable, but the available sources provide a basis for estimating the month, estimate. Otherwise, leave blank. Enter in the format of MM
 

10

Estimated Date

“Yes” if the data was estimated. “No” otherwise.
 

No

GMF Taxonomy Classifications

Taxonomy Details

Known AI Goal Snippets

One or more snippets that justify the classification.
 

(Snippet Text: The large number of dialects in use around the world means that Arabic is particularly difficult for machine translation services to handle, and mistakes are a regular occurrence., Related Classifications: Translation), (Snippet Text: Facebook’s translations are entirely powered by AI, and around 4.5 billion translations are made each day across the social network. , Related Classifications: Translation)

Known AI Goal Classification Discussion

Free text with comments justifying the chosen classification (e.g. based on information on selected snippets and technical analysis), if needed.
 

Translation: Presumably the arabic dialect in the text is not represented adequately in the training data, hence the translation performance issues.

CSETv0 Taxonomy Classifications

Taxonomy Details

Problem Nature

Indicates which, if any, of the following types of AI failure describe the incident: "Specification," i.e. the system's behavior did not align with the true intentions of its designer, operator, etc; "Robustness," i.e. the system operated unsafely because of features or changes in its environment, or in the inputs the system received; "Assurance," i.e. the system could not be adequately monitored or controlled during operation.
 

Specification

Physical System

Where relevant, indicates whether the AI system(s) was embedded into or tightly associated with specific types of hardware.
 

Software only

Level of Autonomy

The degree to which the AI system(s) functions independently from human intervention. "High" means there is no human involved in the system action execution; "Medium" means the system generates a decision and a human oversees the resulting action; "low" means the system generates decision-support output and a human makes a decision and executes an action.
 

High

Nature of End User

"Expert" if users with special training or technical expertise were the ones meant to benefit from the AI system(s)’ operation; "Amateur" if the AI systems were primarily meant to benefit the general public or untrained users.
 

Amateur

Public Sector Deployment

"Yes" if the AI system(s) involved in the accident were being used by the public sector or for the administration of public goods (for example, public transportation). "No" if the system(s) were being used in the private sector or for commercial purposes (for example, a ride-sharing company), on the other.
 

No

Data Inputs

A brief description of the data that the AI system(s) used or were trained on.
 

User posts/input

MIT Taxonomy Classifications

Machine-Classified
Taxonomy Details

Risk Subdomain

A further 23 subdomains create an accessible and understandable classification of hazards and harms associated with AI
 

7.3. Lack of capability or robustness

Risk Domain

The Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental harms, and (7) AI system safety, failures & limitations.
 
  1. AI system safety, failures, and limitations

Entity

Which, if any, entity is presented as the main cause of the risk
 

AI

Timing

The stage in the AI lifecycle at which the risk is presented as occurring
 

Post-deployment

Intent

Whether the risk is presented as occurring as an expected or unexpected outcome from pursuing a goal
 

Unintentional

Incident Reports

Reports Timeline

+21
Facebook auto-translation AI error results in Palestinian man's arrest
Palestinian Arrested After Facebook Translation ErrorHow a Facebook automatic translation error got an innocent man arrestedMan arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’Mixed Messages? The Limits of Automated Social Media Content AnalysisA Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)
Facebook auto-translation AI error results in Palestinian man's arrest

Facebook auto-translation AI error results in Palestinian man's arrest

privacyinternational.org

Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them'

Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them'

haaretz.com

Israeli police arrest Palestinian over Facebook post

Israeli police arrest Palestinian over Facebook post

dailymail.co.uk

Israel police mistakenly arrest Palestinian man for writing ‘good morning’ on Facebook

Israel police mistakenly arrest Palestinian man for writing ‘good morning’ on Facebook

independent.co.uk

Man arrested because of a Facebook mistranslation

Man arrested because of a Facebook mistranslation

alphr.com

Palestinian arrested for saying ‘good morning’ on Facebook

Palestinian arrested for saying ‘good morning’ on Facebook

postcourier.com.pg

Palestinian arrested by Israeli police after Facebook mistranslated his 'good morning' post

Palestinian arrested by Israeli police after Facebook mistranslated his 'good morning' post

ibtimes.co.uk

Israel: Botched Facebook translation lands man in lock-up; ‘Good Morning’ turns ‘Attack Them’

Israel: Botched Facebook translation lands man in lock-up; ‘Good Morning’ turns ‘Attack Them’

newsx.com

Man Arrested After Facebook Translates Turns Greeting Into Threat

Man Arrested After Facebook Translates Turns Greeting Into Threat

digitaltrends.com

Israel arrested a Palestinian after Facebook translated “Good morning” as “Attack them”

Israel arrested a Palestinian after Facebook translated “Good morning” as “Attack them”

qz.com

Facebook translates 'good morning' into 'attack them', leading to arrest

Facebook translates 'good morning' into 'attack them', leading to arrest

theguardian.com

Facebook apologizes after wrong translation sees Palestinian man arrested for posting ‘good morning’

Facebook apologizes after wrong translation sees Palestinian man arrested for posting ‘good morning’

theverge.com

Palestinian Man Arrested After Facebook Auto-Translates 'Good Morning' As 'Attack Them'

Palestinian Man Arrested After Facebook Auto-Translates 'Good Morning' As 'Attack Them'

gizmodo.com.au

Man Arrested After Facebook Translation Error

Man Arrested After Facebook Translation Error

swordandscale.com

Facebook's auto translation AI fail leads to a nightmare for a Palestinian man

Facebook's auto translation AI fail leads to a nightmare for a Palestinian man

mashable.com

Facebook apologizes after wrong translation sees Palestinian man arrested

Facebook apologizes after wrong translation sees Palestinian man arrested

news.ycombinator.com

Man arrested for saying 'good morning' because online translator made a mistake

Man arrested for saying 'good morning' because online translator made a mistake

metro.co.uk

A ‘Good Morning’ Reportedly Gets Man Arrested After Facebook Translate Error

A ‘Good Morning’ Reportedly Gets Man Arrested After Facebook Translate Error

supernewsworld.com

Facebook Apologizes After Translation Fail Leads to False Arrest

Facebook Apologizes After Translation Fail Leads to False Arrest

dailydot.com

This Man Was Arrested For Posting 'Good Morning' On Facebook

This Man Was Arrested For Posting 'Good Morning' On Facebook

thealternativedaily.com

Faulty Facebook translation leads to Palestinian man’s arrest

Faulty Facebook translation leads to Palestinian man’s arrest

tomedes.com

Palestinian Arrested After Facebook Translation Error

Palestinian Arrested After Facebook Translation Error

forward.com

How a Facebook automatic translation error got an innocent man arrested

How a Facebook automatic translation error got an innocent man arrested

translateplus.com

Man arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’

Man arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’

forum.facepunch.com

Mixed Messages? The Limits of Automated Social Media Content Analysis

Mixed Messages? The Limits of Automated Social Media Content Analysis

cdt.org

A Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)

A Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)

finance.yahoo.com

Facebook auto-translation AI error results in Palestinian man's arrest
privacyinternational.org · 2017

A mistake in Facebook's machine translation service led to the arrest and questioning of a Palestinian man by Israeli police. The man, a construction worker on the West Bank, posted a picture of himself leaning against a bulldozer like thos…

Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them'
haaretz.com · 2017

The Israel Police mistakenly arrested a Palestinian worker last week because they relied on automatic translation software to translate a post he wrote on his Facebook page. The Palestinian was arrested after writing “good morning,” which w…

Israeli police arrest Palestinian over Facebook post
dailymail.co.uk · 2017

Israeli police mistakenly arrested a Palestinian who posted 'good morning' in Arabic online which Facebook wrongly translated as 'attack them'.

The man uploaded a picture of himself leaning against a bulldozer at the Israeli settlement of B…

Israel police mistakenly arrest Palestinian man for writing ‘good morning’ on Facebook
independent.co.uk · 2017

A Palestinian man who posted “good morning” in Arabic on Facebook was arrested after Israeli police mistakenly translated his message as “attack them”.

The man, who has not been named, posted a picture of himself working on a construction s…

Man arrested because of a Facebook mistranslation
alphr.com · 2017

A Palestinian man was arrested by Israeli police last week, after a post he left on Facebook saying "good morning" in Arabic was taken to meaning "attack them" in Hebrew.

A spokeswoman for the Israel Police’s West Bank district confirmed to…

Palestinian arrested for saying ‘good morning’ on Facebook
postcourier.com.pg · 2017

Following a goof up by Facebook, a Palestinian man was arrested by the Israeli police for posting ‘Good Morning’ on the social media site.

According to reports, the man, a construction worker said “Good Morning” in Arabic in Facebook post, …

Palestinian arrested by Israeli police after Facebook mistranslated his 'good morning' post
ibtimes.co.uk · 2017

A Palestinian man in the West Bank was arrested by Israeli police after he wrote "good morning" in Arabic in a Facebook post that was mistranslated by the company's automatic translation software. Construction worker Halawim Halawi had repo…

Israel: Botched Facebook translation lands man in lock-up; ‘Good Morning’ turns ‘Attack Them’
newsx.com · 2017

Palestinian builder Halawim Halawi who posted 'Good Morning' on his Facebook account which awarded him a few hours in the lockup. Facebook's automatic translation software which interpreted the post to mean "attack them" in Hebrew and "hurt…

Man Arrested After Facebook Translates Turns Greeting Into Threat
digitaltrends.com · 2017

Share

Facebook’s auto-translate feature allows users to connect beyond language barriers but one incorrect translation of a simple “good morning” proved to be a bit of a disaster for one Palestinian man. When the construction worker posted …

Israel arrested a Palestinian after Facebook translated “Good morning” as “Attack them”
qz.com · 2017

Facebook says it automatically translates about 4.5 billion posts each day. With new research and advancements, the technology keeps improving, but it’s not perfect—and its flaws can have serious consequences.

Last week, a Palestinian man w…

Facebook translates 'good morning' into 'attack them', leading to arrest
theguardian.com · 2017

Palestinian man questioned by Israeli police after embarrassing mistranslation of caption under photo of him leaning against bulldozer

Facebook has apologised after an error in its machine-translation service saw Israeli police arrest a Pal…

Facebook apologizes after wrong translation sees Palestinian man arrested for posting ‘good morning’
theverge.com · 2017

Facebook has apologized after a Palestinian man was arrested by Israeli police for a post saying “good morning” that its automatic-translation service erroneously translated as “attack them” in Hebrew and “hurt them” in English, reports Isr…

Palestinian Man Arrested After Facebook Auto-Translates 'Good Morning' As 'Attack Them'
gizmodo.com.au · 2017

A Palestinian construction worker was arrested by Israeli police after Facebook incorrectly translated the text of one of his posts. Haaretz reports that the man uploaded a picture from his job at a construction site with the text "good mor…

Man Arrested After Facebook Translation Error
swordandscale.com · 2017

They say a picture is worth a thousand words, but it only took two of them for a Palestinian man to find himself being placed under arrest after a Facebook translation blunder.

A construction worker for the West Bank settlement of Beitar Il…

Facebook's auto translation AI fail leads to a nightmare for a Palestinian man
mashable.com · 2017

Bad translations often make good comedy, but in the case of this Palestinian man, they marked the start of a nightmare.

Israeli police arrested the man, a construction worker in a West Bank settlement near Jerusalem, after he posted on Face…

Facebook apologizes after wrong translation sees Palestinian man arrested
news.ycombinator.com · 2017

Hacker news discussion of Facebook's apology.

Man arrested for saying 'good morning' because online translator made a mistake
metro.co.uk · 2017

A construction worker was arrested after he posted ‘Good morning’ on Facebook, because the automatic translation got it wrong.

The Palestinian man had posted a photo at work with a cigarette and a coffee, adding a greeting to his followers …

A ‘Good Morning’ Reportedly Gets Man Arrested After Facebook Translate Error
supernewsworld.com · 2017

Digital Trends has reported that Facebook’s auto-translate feature allows users to connect beyond language barriers but one incorrect translation of a simple “good morning” proved to be a bit of a disaster for one Palestinian man. When the …

Facebook Apologizes After Translation Fail Leads to False Arrest
dailydot.com · 2017

No translation service is perfect, but you wouldn’t expect it to get you arrested.

Facebook admitted that an error in its artificial intelligence-based auto-translation feature led to the arrest of a Palestinian man who posted “good morning…

This Man Was Arrested For Posting 'Good Morning' On Facebook
thealternativedaily.com · 2017

Israeli police accidentally arrested a Palestinian man after he posted “Good morning” in Arabic on Facebook, together with a photo of himself at work.

The man was working at a building site on a West Bank settlement near Jerusalem and posed…

Faulty Facebook translation leads to Palestinian man’s arrest
tomedes.com · 2017

It was only last week that we discussed the flaws of machine translation. In that case, the focus of the article was on the WeChat messaging app’s racial slurs. Now, a new story has once more highlighted the dangers posed by the machines th…

Palestinian Arrested After Facebook Translation Error
forward.com · 2017

A Palestinian man was mistakenly arrested in Israel after Facebook’s translation software turned his “good morning” photo caption into a threat.

The man, who works in a West Bank settlement, posted a photo of himself next to a bulldozer wit…

How a Facebook automatic translation error got an innocent man arrested
translateplus.com · 2017

Some automatic translation caused mistakes are harmless enough; others not so much. The occasional language faux-pas can be good for a laugh but you never know when a seemingly harmless mishap can cause real damage. The latest example of tr…

Man arrested after ‘good morning’ post was mistranslated by Facebook as ‘attack them’
forum.facepunch.com · 2018

[QUOTE]Israeli police arrested a Palestinian man after his “good morning” post was translated by Facebook as “attack them.” Social media monitoring combined with Facebook’s AI mistranslating a post led to a man being wrongfully arrested for…

Mixed Messages? The Limits of Automated Social Media Content Analysis
cdt.org · 2018

Governments and companies are turning to automated tools to make sense of what people post on social media. Policymakers routinely call for social media companies to identify and take down hate speech, terrorist propaganda, harassment, fake…

A Google translation isn't enough evidence to send someone to jail, finds the judge in a narcotics case (GOOG, GOOGL)
finance.yahoo.com · 2018

Greg Sandoval/Business Insider

A man from Mexico was pulled over on the highway in Kansas and was suspected of carrying drugs.

With the help of Google Translate, the Kansas Highway trooper asked the man in Spanish if he can search his car. …

Variants

A "variant" is an incident that shares the same causative factors, produces similar harms, and involves the same intelligent systems as a known AI incident. Rather than index variants as entirely separate incidents, we list variations of incidents under the first similar incident submitted to the database. Unlike other submission types to the incident database, variants are not required to have reporting in evidence external to the Incident Database. Learn more from the research paper.

Similar Incidents

By textual similarity

Did our AI mess up? Flag the unrelated incidents

Images of Black People Labeled as Gorillas

Images of Black People Labeled as Gorillas

Jun 2015 · 24 reports
Biased Google Image Results

Biased Google Image Results

Mar 2016 · 18 reports
Security Robot Drowns Itself in a Fountain

Security Robot Drowns Itself in a Fountain

Jul 2017 · 30 reports
Previous IncidentNext Incident

Similar Incidents

By textual similarity

Did our AI mess up? Flag the unrelated incidents

Images of Black People Labeled as Gorillas

Images of Black People Labeled as Gorillas

Jun 2015 · 24 reports
Biased Google Image Results

Biased Google Image Results

Mar 2016 · 18 reports
Security Robot Drowns Itself in a Fountain

Security Robot Drowns Itself in a Fountain

Jul 2017 · 30 reports

Research

  • Defining an “AI Incident”
  • Defining an “AI Incident Response”
  • Database Roadmap
  • Related Work
  • Download Complete Database

Project and Community

  • About
  • Contact and Follow
  • Apps and Summaries
  • Editor’s Guide

Incidents

  • All Incidents in List Form
  • Flagged Incidents
  • Submission Queue
  • Classifications View
  • Taxonomies

2023 - AI Incident Database

  • Terms of use
  • Privacy Policy
  • Open twitterOpen githubOpen rssOpen facebookOpen linkedin
  • 9d70fba