NBC debuts AI-generated Jim Fagan voice for NBA coverage (awfulannouncing.com)

NBC quietly rolled out an AI-generated version of late announcer Jim Fagan’s voice during Peacock’s preseason broadcast of Bulls vs. Cavaliers, using what the network described as “AI voice synthesis technology” to run a halftime promo. Fagan, who died in 2017 and hadn’t narrated NBA promos since the early 2000s, was recreated in a way that many viewers found unsettling: familiar but noticeably synthetic. NBC has used synthetic voices before (notably Al Michaels for personalized Olympic highlights in 2024), but the key difference here is that Fagan is deceased, which magnifies the ethical and emotional reaction. For the AI/ML community this is a high-profile example of both capability and convolution: modern speaker-synthesis models can convincingly mimic archived voices by training on existing broadcasts, matching timbre and prosody well enough for broadcast promos, yet still produce uncanny artifacts. The incident spotlights urgent technical and policy needs—provenance and detectable watermarks in synthetic audio, consent and posthumous rights handling, dataset transparency, and robustness against misuse (deepfakes). It’s also a reminder that practical deployment requires social and legal guardrails as much as signal quality improvements: just because a voice can be synthesized doesn’t mean it should be used without clear ethical, legal, and audience-consent frameworks.