Ethicality AI Use Standard
EAUS
EAUS is a practical standard for transparent, human-led business use of AI. It focuses on three things: being open about meaningful AI use, keeping people responsible for important work and decisions, and protecting confidential information.
Three ideas
Be transparent
Say so when AI meaningfully contributes to your work, and be accurate about how you use it.
Important decisions stay with people
A person reviews material AI-assisted work and approves important decisions. AI does not make final consequential decisions about people on its own.
Protect confidential information
Confidential, privileged and sensitive information stays out of public or unapproved AI tools.
The standard
Fifteen requirements, in three themes.
01
Simple rules and confidential information
Employees know how they may use AI, which tools are allowed, and what must never be typed into one.
R1
The organization has simple written rules for AI use that employees who use AI can access.
R2
Employees are told that confidential, privileged or sensitive business information must not be entered into public or unapproved AI tools.
R3
Passwords, security credentials and similarly sensitive access information are never entered into public or unapproved AI tools.
R4
Employees know which AI tools they are permitted to use for work.
R15
Employees who use AI know the basic disclosure, confidentiality and human-review rules and know who to ask if they are unsure.
02
Openness about AI use
Clients and people know when AI meaningfully contributed, and public statements are accurate.
R5
The organization is open with clients or customers about meaningful AI use in its work.
R6
When AI substantially contributes to client-facing work, the organization discloses that use where a reasonable client would expect to know.
R7
People are told when they are directly interacting with an AI system, chatbot or agent rather than a person.
R8
If AI plays a meaningful role in a consequential decision about a person, that person is told that AI was involved.
R9
Public statements about the organization's use of AI are accurate and not misleading.
03
Human responsibility
A person reviews material AI-assisted work, approves important decisions, and stays responsible for the outcome.
R10
Material AI-assisted work is reviewed by a person before it is sent, published or relied upon externally.
R11
Important recommendations, advice or decisions informed by AI receive appropriate human approval before action is taken.
R12
AI does not make final consequential decisions about people on its own.
R13
The person reviewing AI-assisted work has the authority to change, reject or disregard the AI output.
R14
A person, not the AI system, remains responsible for the final work or decision.
Need a starting point? See ready-to-use templates and examples for these requirements.
How answers work
Yes, No, or Not applicable.
Yes
The organization already follows the practice described.
No
The practice is not yet in place. We tell you which requirement has not yet been met and what is missing. Where relevant, a free template can help.
Not applicable
The requirement genuinely does not apply to the organization's work.
An organization either meets the applicable EAUS requirements, or it is told exactly what to change before it can certify.
The outcome
Our badge.
Ethicality Badge
One badge. An organization either meets the applicable EAUS requirements or we explain which applicable requirements have not yet been met.
An organization that does not yet meet the applicable requirements receives private findings and can return once the gaps are closed. Nothing is published unless certification is issued, and every displayed badge links to the holder's public registry entry.
Keeping a certification
What holders commit to for the year.
A five-minute annual re-attestation against the same fifteen statements.
Keep the certified practices in place and answer truthfully.
Tell us if your organization stops materially meeting the standard.
Ethicality may request information or suspend certification where there is a credible concern that the representations were inaccurate.
How this standard changes
Reviewed once a year, with the version on the record.
EAUS is reviewed annually and updated when needed. Every version carries a number, every certificate and registry entry records the version it was issued under, and a certification already issued is never reassessed against a later version during its term. Renewal and re-certification are assessed against the version current at that time, and where a new version adds requirements we say so when we publish it so holders know what changed before renewal.
Key terms
The language this standard uses.
Human responsibility, meaningful AI contribution, confidential information, and re-attestation are all defined in one place.
Apply now
Start your certification.
A quick application and you're on your way to Ethicality certification.
Ethicality is a private certification program. An Ethicality Badge reflects an independent assessment of the AI-use practices an organization disclosed to us, against the published EAUS standard, as at the assessment date. It does not certify compliance with any law.