Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is Non-Repudiation?

Non-repudiation is the assurance that whoever sent a message or approved a transaction cannot later deny it, backed by evidence strong enough to stand up to challenge.

Illustration for the cybersecurity term Non-Repudiation
Home / Blog / Cybersecurity Terms / What Is Non-Repudiation?
17 November 2026

What Is Non-Repudiation?

What Is Non-Repudiation?

Non-repudiation is the property that stops someone credibly denying an action they took. It is not about keeping a secret or blocking an attacker; it is about being able to prove, later and to a sceptical third party, that a particular party sent this message, approved that payment, or deleted that record. Where confidentiality and integrity protect the data, non-repudiation protects the account of what happened.

Why Logging Alone Is Not Enough

Most teams assume their audit trail provides this. It usually does not. A log line saying user=jsmith action=approve amount=48000 proves only that something wrote that line. If ten people know the shared service account password, if administrators can edit the log file, or if the clock is unsynchronised, the entry is an assertion rather than evidence.

Three weaknesses recur:

The Mechanisms That Do Work

Digital signatures are the strongest building block. The signer holds a private key that nobody else can use, and anyone with the matching certificate can verify the signature over the exact bytes signed. Because the key is not shared, verification links the document to the key holder. This depends entirely on key custody: a private key on a smart card or hardware security module supports a far stronger claim than one sitting in a user profile. See our entry on public key infrastructure for how the certificates and revocation behind this are managed.

Trusted timestamps add the when. A time-stamping authority following RFC 3161 signs a hash of your document together with its own clock reading, so the document provably existed before that moment. This is what keeps a signature verifiable after the signing certificate expires.

Tamper-evident logging adds the what happened next. Writing log entries to append-only or WORM storage, chaining each entry’s hash to the previous one, and shipping events off the originating host within seconds all make silent alteration detectable. Regular hashing of archived logs, with the hashes held separately, gives you something to compare against.

Strong authentication underpins all of it. Multi-factor authentication with phishing-resistant factors makes the claim that someone else used the account much harder to sustain.

Where It Appears in Security Models

Non-repudiation is the R in STRIDE: repudiation is the threat, non-repudiation the property it violates. ISO/IEC 27000 lists it alongside authenticity and accountability as a property that information security may need to preserve, beyond the core CIA triad. In digital forensics, chain of custody is non-repudiation applied to evidence handling: every transfer is signed, timed, and witnessed so that defence counsel cannot argue the exhibit was altered.

Legal Standing in the UK

Section 7 of the Electronic Communications Act 2000 makes an electronic signature, and the certification of it, admissible in evidence in legal proceedings about the authenticity or integrity of a communication. The UK’s trust services regime, derived from the eIDAS Regulation as retained after EU exit and supervised by the Information Commissioner’s Office, recognises advanced and qualified electronic signatures, with qualified signatures carrying the strongest presumption. In 2019 the Law Commission confirmed that electronic signatures are capable of executing most documents under English law.

Admissibility is not the same as weight. A court will look at how the key was protected, whether the signatory understood what they were signing, and whether the audit trail is internally consistent. A clickwrap tick box with a timestamp in an editable database is admissible and weak; a qualified signature made on a hardware token is admissible and hard to argue with.

Designing for It

Decide which actions in your systems must survive a dispute—payment approval, contract execution, prescription issue, evidence handling, privileged configuration change—and apply the strong mechanisms there rather than everywhere. For everything else, accountable logging is proportionate. Record the design decision, because the question always arrives after the event rather than before it. Our digital forensic and incident response services work with the evidence organisations actually have, and the gaps are rarely in the technology they bought.

FAQ – Frequently Asked Questions

Is a Scanned Signature on a PDF Good Enough?

Rarely. A pasted image proves nothing about who placed it there and can be copied from any other document. A cryptographic signature applied with a private key the signer controls, ideally with a trusted timestamp, is what links the document to the person. If a dispute is plausible, use a qualified or advanced electronic signature service rather than an image.

Can Non-Repudiation Ever Be Absolute?

No. Every mechanism rests on the assumption that the signing key or credential was under the holder’s sole control. A stolen smart card, malware on the signing device, or a compromised certificate authority all give a signatory a genuine basis for denial. Non-repudiation raises the evidential bar substantially; it does not make denial impossible.

How Do Service Accounts Affect Non-Repudiation?

They undermine it. Any credential shared between people or embedded in scripts means the log shows the account, not the human behind it. Where shared access is unavoidable, route it through a privileged access management system that authenticates the individual first, brokers the session, and records it, so the audit trail still names a person.

What Should Our Logs Capture to Support It?

Who performed the action, what changed before and after, the authenticated session and source address, and an accurate synchronised timestamp. Send events off the originating host immediately, store them on append-only media, and restrict who can alter retention settings. Without those last two controls, an administrator or an attacker with administrative rights can rewrite the record.

Related Terms

You May Be Interested In These

What Is Post-Quantum Cryptography?

Cybersecurity Terms

17 November 2026

What Is Post-Quantum Cryptography?

Read More
What Is Hypervisor Security?

Cybersecurity Terms

17 November 2026

What Is Hypervisor Security?

Read More
What Is Crypto Exchange Security?

Cybersecurity Terms

17 November 2026

What Is Crypto Exchange Security?

Read More