False Positive
A false positive is a security alert that flags something harmless as malicious, such as a legitimate program blocked as malware. A false negative is the opposite and more dangerous: a real attack that no tool flags.
Also called: false alarm, false negative, true positive
What is a false positive?
Security tools judge activity as good or bad, and they sometimes get it wrong. A false positive is a false alarm: the tool says there's a threat, but there isn't. NIST defines it as an alert that incorrectly indicates a vulnerability is present.
There are four possible outcomes:
- True positive: a real threat, correctly flagged.
- False positive: no threat, but the tool raises an alert.
- True negative: no threat, and the tool stays quiet.
- False negative: a real threat that the tool misses.
Common examples
- Antivirus quarantining a legitimate accounting or line-of-business program
- An email filter sending a real customer invoice to spam
- A vulnerability scanner reporting a flaw that was already patched, because it read the software's version number but not the fix
- An "impossible travel" sign-in alert caused by a VPN
- An intrusion detection system flagging a backup job as data exfiltration
Why they matter
A few false positives are normal. Too many cause alert fatigue: people stop trusting alerts and start ignoring them, and a real attack slips through. False negatives are worse, because nobody knows anything happened. Tuning a tool to cut false positives can increase false negatives, so it's a balance.
How to reduce false positives
- Tune rules to your environment, for example by allowing known business software and scheduled jobs.
- Validate scan findings manually or with authenticated scans, as NIST's testing guide recommends, before assigning fixes.
- Add context, such as asset importance and user role, so tools can tell normal from unusual.
- Review exceptions regularly so old allowances don't become blind spots.
A good managed security provider filters false positives before they reach you.
Related terms
Sources
Drafted with AI assistance, then edited and fact-checked by a person. How we use AI.
