It depends on what the tool actually does, and the dividing line is emotion. A video interview is not banned in itself; inferring a candidate's emotional state is.
On the prohibited side: Article 5(1)(f) bans AI that infers a person's emotions in the workplace, and the Commission's guidelines extend “workplace” to recruitment. A tool that scores “confidence”, “enthusiasm”, or “cultural fit” from facial expression, voice tone, or body language is caught, and has been banned since February 2025, not deferred to 2027.
On the compliant side: a tool that records and transcribes answers, or scores responses against role-relevant competencies, without inferring emotional or affective state, is permitted. It is still high-risk under §4 as candidate evaluation, so it needs human oversight, candidate transparency, and bias testing, but it is lawful. The Scorecard helps you work out which side of that line a given tool sits on.