Incident 137: Israeli Tax Authority Employed Opaque Algorithm to Impose Fines, Reportedly Refusing to Provide an Explanation for Amount Calculation to a Farmer

Description: An Israeli farmer was imposed a computer generated fine by the tax authority, who allegedly were not able to explain its calculation, and refused to disclose the program and its source code.

Tools

New Report New Response DiscoverView History

Entities

View all entities

Alleged: Israeli Tax Authority developed and deployed an AI system, which harmed Moshe Har Shemesh and Israeli people having tax fines.

Incident Stats

Incident ID

137

Report Count

Incident Date

2021-01-11

Editors

Sean McGregor, Khoa Lam

Applied Taxonomies

GMF, CSETv1, MIT

CSETv1 Taxonomy Classifications

Taxonomy Details

Incident Number

137

GMF Taxonomy Classifications

Taxonomy Details

Known AI Goal Snippets

(Snippet Text: The story began in 2014 when the farm Har Shemesh, which is not part of any community, asked the Tax Authority to explain how it had calculated a fine they were required to pay., Related Classifications: Financial Processing)

Known AI Goal Classification Discussion

Financial Processing: The system probably does not use AI at all.

MIT Taxonomy Classifications

Machine-Classified

Taxonomy Details

Risk Subdomain

7.4. Lack of transparency or interpretability

Risk Domain

AI system safety, failures, and limitations

Entity

Timing

Post-deployment

Intent

Intentional

Incident Reports

Reports Timeline

When an Israeli Farmer Declared War on an Algorithm

haaretz.com

haaretz.com · 2021

A tax dispute between a ranch in Israel’s southern Negev desert and the country’s tax authority has given rise to an issue which could have far reaching repercussions on the public’s ability to oversee and even understand government acts.

T…

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.