Trustworthy AI

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

The property of an AI system that earns justified trust: reliable, safe, fair, transparent, and accountable, as set out in the EU's seven requirements.

What is Trustworthy AI?

Trustworthy AI is the property of an AI system that justifies the trust people place in it: it performs reliably, resists misuse, treats people fairly, protects their data, and can be explained and held to account.

In Europe the term has a precise pedigree. The EU’s High-Level Expert Group on AI published its Ethics Guidelines for Trustworthy AI in 2019, and that framework shaped the EU AI Act.

How Trustworthy AI Works

The EU guidelines hold that trustworthy AI must be lawful, ethical, and technically sound, and they set seven requirements a system should meet:

  1. Human agency and oversight
  2. Technical robustness and safety
  3. Privacy and data governance
  4. Transparency
  5. Diversity, non-discrimination and fairness
  6. Societal and environmental well-being
  7. Accountability

The framing matters because these are properties of the system, checkable against evidence: test results, documentation, logs, audit findings. Trustworthiness is not a claim a vendor makes; it is a state a system either demonstrates or does not.

Trustworthy AI vs Responsible AI

Trustworthy AI is what a system is; responsible AI is what an organization does to produce it. One names the destination, the other the route.

Both draw their content from ethics in AI, the field that works out which properties are worth demanding in the first place. When a regulator inspects a deployed system, it evaluates trustworthiness; when it inspects the provider’s processes, it evaluates responsibility.

Trustworthy AIResponsible AI
What it namesWhat the resulting system isHow an organization behaves
Kind of thingA property you verifyA process you follow
EvidenceTest results, documentation, logs, audit findingsPolicies, review gates, named owners
A regulator inspectsThe deployed systemThe provider’s processes

Example of Trustworthy AI

The EU AI Act, in force since August 2024, turned the 2019 guidelines into enforceable obligations. Take a CV-screening tool used in hiring, which the Act classifies as high-risk.

Before placing it on the EU market, the provider must run a risk-management process, meet data governance requirements for its training data, produce technical documentation, keep logs of its operation, design it so a human can oversee and override it, and achieve appropriate levels of accuracy, robustness, and cybersecurity.

Each obligation traces back to one of the seven requirements. What began as a definition of trustworthiness became a checklist a regulator can enforce, and a system that clears the checklist has demonstrated the property rather than asserted it.

Related AI terms: Responsible AI · Ethics in AI · EU AI Act · Robustness · Transparency

Did you like the Trustworthy AI 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