Mitigation

Last Updated: July 29, 2026 | By Mihail Sebastian | AI Dictionary

The actions taken to reduce an identified AI risk – technical fixes, human oversight, and use restrictions – chosen after assessment ranks what could go wrong.

What is Mitigation?

Mitigation is the set of actions taken to reduce an identified AI risk: fixing the system, adding human oversight, restricting how it is used, or deciding not to deploy it at all.

It is one step in a larger cycle. Risk management first identifies what could go wrong and assesses how badly; mitigation is what you do about the risks that ranked worst. Its counterpart is resilience – mitigation reduces risk before harm occurs, resilience determines how the system recovers once something breaks anyway.

Types of Mitigation

Regulation increasingly expects this work to be documented. The EU AI Act requires providers of high-risk AI systems to run a risk management process that includes adopting mitigation measures for the risks it identifies.

Example of Mitigation

A company audits its resume-screening model and finds it scores women lower for technical roles – the same failure pattern that led Amazon to abandon its experimental recruiting tool. Assessment ranks this risk first: it harms applicants and exposes the company to discrimination claims.

The team mitigates on all three fronts. Technically, they rebalance the training data and strip features that act as gender proxies, such as the names of single-sex colleges. Procedurally, every rejection the model is least confident about goes to a human recruiter. In scope, the model is limited to shortlisting; it cannot reject anyone outright.

Before redeployment, they re-run the audit to confirm the score gap has closed, then schedule it quarterly. Mitigation is not a one-time patch but a change they now have to keep proving works.

Related AI terms: Risk Management · AI Risk · Resilience · Bias · Human-in-the-Loop

Did you like the Mitigation gist?

Learn about 250+ need-to-know artificial intelligence terms in the AI Dictionary.

Mihail Sebastian — Writes about AI governance, regulation, and the technology behind them. Placeholder bio — replace with a real credential line. About

Read the Governor's Letter

Stay ahead with Governor's Letter, the newsletter delivering expert insights, AI updates, and curated knowledge directly to your inbox.

By subscribing to the Governor's Letter, you consent to receive emails from AI Guv.
We respect your privacy - read our Privacy Policy to learn how we protect your information.

Browse All AI Terms A–Z

Every term in the dictionary, in alphabetical order. Jump to a letter or scroll the full list.

A

B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z