The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2537
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T07:50:16.768062Z
status: published
lens: p_space
sector: sports
headline: Effects of Human-AI Voice Interaction in Sports Commentary on Perceived Credibility, Emotional Response, and Cognitive Recall
dek: This study examines how commentary source (human vs. artificial intelligence [AI]), commentator gender, and accent shape perceived credibility, affective responses, and cognitive recall in sports broadcasting. A 2 (source) × 2 (commentator gender) × 2 (accent) between-subjects experiment (N = 867) showed that human-voiced commentary received higher ratings than AI-voiced commentary for credibility, audience perception, and positive affect, while eliciting lower negative affect. Male commentator voices received h…
gain_title: (none)
problem_title: In an experiment with 867 participants, AI-voiced sports commentary got lower ratings than human-voiced commentary for credibility, audience perception and positive feeling, and drew more negative feeling.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: In an experiment with 867 participants, AI-voiced sports commentary got lower ratings than human-voiced commentary for credibility, audience perception and positive feeling, and drew more negative feeling.
problem_evidence: human-voiced commentary received higher ratings than AI-voiced commentary for credibility, audience perception, and positive affect, while eliciting lower negative affect
quick_read: An experiment with 867 participants compared human and AI voices delivering sports commentary. Human-voiced commentary scored higher than AI-voiced commentary on credibility, audience perception and positive affect, and produced lower negative affect.

The design crossed three factors: commentary source (human or AI), commentator gender and accent. Male commentator voices received higher credibility and audience perception ratings across conditions, while accent effects were weak.

Recall of the content held steady across gender and accent conditions and differed only minimally by source. The authors read this as a gap between socially shaped judgments of the voice and comparatively stable memory of what was said.
limitation: 
tag: Evidence-backed problem
key_points: The study was a 2 (source) × 2 (commentator gender) × 2 (accent) between-subjects experiment with 867 participants. | Human-voiced commentary was rated higher than AI-voiced commentary for credibility, audience perception and positive affect. | Human-voiced commentary also produced lower negative affect than AI-voiced commentary. | Male commentator voices received higher credibility and audience perception ratings across conditions. | Accent effects were weak. | Cognitive recall stayed stable across gender and accent conditions and differed minimally by source.
rundown: The researchers tested how listeners respond to sports broadcasting when the commentary voice is human or AI, and whether the commentator's gender or accent changes the response. Outcomes were perceived credibility, emotional response and cognitive recall.

Human voices came out ahead on evaluative and emotional measures, while recall of the content barely differed by source. The authors interpret the gap through the Computers Are Social Actors paradigm, social presence theory and symbolic power.

Based on the results, the authors recommend a context-sensitive approach: human voices for emotionally intense commentary and AI voices for structured, information-oriented segments.
sources:
- peer_reviewed | International Journal of Human–Computer Interaction | https://doi.org/10.1080/10447318.2026.2732386 | 2026-09-18
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

b24c94722b3e731c945d27d4524c7223030e03d799ce37a418126c88c7c9836e

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000