What is the difference between speech recognition and voice recognition?

The two terms answer different questions. Speech recognition answers "what was said?" and produces text. Voice recognition answers either that same question or "who is speaking?", depending on who uses the term. Once you know which question a product is answering, the difference is clear.

Speech recognition is about the words, speaker recognition is about the person

Speech recognition is the task of turning spoken words into text. The IEEE Technology Navigator separates it from speaker recognition, which it defines as identifying or verifying a person's identity from the characteristics of their voice ¹. The two tasks use voice as input, but they want different outputs: words for one, and a name or a yes-or-no match for the other.

The same transcription job also goes by several names. Researchers say automatic speech recognition (ASR), and products say speech to text or dictation. The textbook by Jurafsky and Martin calls transcribing speech into writing one of the earliest goals of computer language processing ². ASR and speech to text name the same job. See automatic speech recognition for the research term.

Yes, voice recognition can identify a person, in two ways

Speaker recognition has two tasks, and the difference between them is whether the system is given a claimed identity. In speaker identification, no identity is claimed, and the system searches a database of known speakers for the closest match ³. In speaker verification, the system accepts or rejects a claimed identity by checking whether the voice matches stored models for the person claimed ⁴. So a voice can be used to find who is speaking, or to check that a speaker is who they say they are.

It depends: a voice match is not proof of identity

Recognizing a voice is not the same as proving who is talking. IEEE notes that speaker recognition must contend with spoofing attacks, which can include synthesizing speech with text-to-speech tools ⁵. A recording or a synthetic copy can be used to imitate a person.

Voiceprints also carry legal weight. Illinois's Biometric Information Privacy Act defines biometric identifiers to include voiceprints, alongside fingerprints and iris scans ⁶. Under that definition, a voiceprint is a biometric identifier in Illinois. This describes the statute, not legal advice for your situation.

What to do

1. Ask which task a product means. A dictation app performs speech recognition. A login by voice performs speaker verification, and the two need different checks.
2. Treat a voice match as one weak factor, not as proof of identity. Do not rely on it alone to protect an account or a payment.
3. For what was said, read our guide to how speech recognition works. For the wider subject, see the overview of speech recognition.

Related: Is speech recognition AI?, Is speech recognition part of NLP? and What is speech recognition used for?. For the term, see speaker diarization, which answers who spoke when in a recording.