Kan AI slå trænede mennesker i læbebevægelseslæsning ?
Afgiv din stemme — læs så hvad vores redaktør og AI-modellerne fandt.
DeepMind viste dette i 2022 med en transformer-baseret model, der overgik professionelle læbere på TV-nyhedsindslag.
Background
Researchers have made significant progress in developing artificial intelligence systems that can lip-read, with some studies demonstrating that AI models can outperform trained human lip-readers in certain conditions. These AI systems use computer vision and machine learning algorithms to analyze the movements of a person's lips and identify the corresponding speech sounds. While the accuracy of AI lip-reading systems can vary depending on factors such as the quality of the video input and the complexity of the speech, they have shown promising results in various experiments. Overall, the current state of the art in AI lip-reading suggests that these systems can indeed beat trained humans in certain scenarios.
— Enriched May 9, 2026 · Source: University of Oxford
Foreslå et tag
Mangler et begreb i dette emne? Foreslå det, admin gennemgår.
Status senest tjekket August 14, 2026.
Galleri
Kan AI slå trænede mennesker i læbebevægelseslæsning?
Juryen fandt et klart bekræftende svar.
Efter livlig drøftelse fandt juryen sig med to stemmer mod én for at anerkende AI's overlegenhed i læbestavning, hvor den tøvende jurymedlem insisterede på, at teknologien fortsat er ujævn på tværs af virkelige forhold. Flertallet imødegik dette med, at nutidens modeller allerede i standardiserede tests og laboratorieindstillinger overgår endog erfarne menneskelige tolke. Dom: "Stumfilm kan nu læses, og maskinerne vinder pladserne i første række."
After lively deliberation, the jury found itself two votes to one in favor of recognizing AI supremacy in lip-reading, with the hesitant juror insisting the technology remains uneven across real-world conditions. The majority countered that on standardized tests and in laboratory settings, today’s models already outperform even seasoned human interpreters. Ruling: “Silent movies can now be read, and the machines win the front-row seats.”
But the data is real.
The Case File
Across 20 sessions, 51 jurors have heard this case. Combined tally: 24 YES · 25 ALMOST · 2 NO · 0 IN RESEARCH.
Note: cumulative includes older juror opinions. The current session tally above is the live verdict.
By a vote of 2 — 1 — 0, the panel returns a verdict of JA, with verdict confidence of 88%. The court so orders. Verdict upgraded from prior session.
"Lip-reading AI exists but accuracy varies"
"AI lip-reading models (e.g., LipNet, AV-HuBERT, Wav2Lip) exceed human performance on benchmark datasets like LRW and LRS3."
"AI systems have demonstrated lip-reading accuracy exceeding that of professional human lip-readers in controlled environments and on specific datasets."
Individuelle nævningers udtalelser vises på originalengelsk for at bevare bevismæssig præcision.
Hvad publikum mener
Nej 6% · Ja 75% · Måske 19% 150 votesDiskussion
no comments⚖ 20 jury checks · seneste for 5 dage siden
Hver række er et separat jurytjek. Nævninger er AI-modeller (identiteter holdt neutrale med vilje). Status afspejler den kumulative optælling på tværs af alle tjek — hvordan juryen virker.