Elevenlabs poorly dubbing videos compilation
Credibility score: 46/100 — Mixed Credibility. Several questionable claims detected. Watch with healthy skepticism.
BSmeter analyzed "Elevenlabs poorly dubbing videos compilation" and rated it 46/100 for credibility (a BS score of 54/100 — mixed credibility), on 2026-08-26. Its weakest claim — "Shows AI switching between languages, technical glitches, and random screaming — complete system breakdown" — scored 10/100 and was flagged as confidence mismatch. 38 claims were checked against the video transcript. Scores are produced by BSmeter's AI analysis of the transcript, not independent human verification.
Of 38 claims analyzed: 19 scored under 40, 3 between 40 and 69, and 16 at 70 or above.
Claims analyzed
Video opens with a highlight reel preview — Just Vibes (50/100)
At 0:01
Teaser montage of ElevenLabs audio fails — sets up the punchline before the evidence arrives.
Why this score: The first 30 seconds is a rapid montage of broken dubbing clips. This isn't a claim; it's a preview designed to prime you for the 'bad AI' frame. The technique is emotional button — you're already laughing before you hear any context about what ElevenLabs actually promised.
Original quote: “[0:01] 100 years, 150 years. [0:05] MYSELF [0:07] THIS is the English translation for [0:21] Yes, 155 years. Make sure it is old. [0:25] I have a great affection for this country... [0:28] Whereas I am the father of parties and roundabouts to do it with this grip, it will make you make a No, I'm not”
Random vulgar phrases inserted mid-dub — classic AI hallucination — Missing Context (20/100)
At 2:34
ElevenLabs drops completely unrelated profanity instead of translating — zero context preserved
Why this score: The model has no memory of the original dialogue or scene, so it defaults to statistically common but wildly inappropriate phrases. This isn't just bad lip-sync; it's a total semantic collapse where the AI fabricates content from nothing. A neutral framing would be 'AI-generated nonsense' instead of pretending it's a translation.
Original quote: “Pull my penis. I am gay. [music] Strange things are happening to me. Suck my balls.”
AI mangles dialogue into nonsense words — No Frame (75/100)
At 4:51
Shows ElevenLabs turning clear speech into gibberish — proof of concept right here.
Why this score: The transcript presents the actual output of ElevenLabs dubbing, demonstrating the technology's failure to preserve original meaning. This is direct evidence rather than rhetorical framing.
Original quote: “Ah. The ashes on you. It is not by fuel other than um Jago”
AI creates absurdly inappropriate mistranslations — No Frame (75/100)
At 5:15
From normal dialogue to 'musical plays about suicide' — shows how completely the meaning gets lost.
Why this score: The dubbed output completely misrepresents the original content, transforming innocuous speech into disturbing subject matter. This demonstrates the risk of AI dubbing creating entirely wrong narratives.
Original quote: “The organization of musical plays about suicide. It is going to be It is truly going Don't correct the program.”
AI refuses to translate content it deems problematic — No Frame (75/100)
At 5:38
ElevenLabs blocking translation entirely — another failure mode exposed.
Why this score: The system sometimes refuses to process content rather than attempting a flawed translation, revealing limitations in its content filtering and processing capabilities.
Original quote: “I'm sorry, but I can't translate that.”
AI switches between dramatic medieval speech and technical error messages — No Frame (75/100)
At 5:46
Mixing epic fantasy dialogue with AI error handling — shows the system breaking character completely.
Why this score: The dubbed output jumps from dramatic, stylized language to a generic technical correction message, demonstrating the AI's inability to maintain consistent tone or context across a translation.
Original quote: “Yakov has dispatched a message. It is imperative that the traitors be put to death so as to avert the opening of the Garth's gates. The sentence you provided appears to have a typo or is confusing.”
AI produces repetitive nonsense then breaks down emotionally — No Frame (75/100)
At 6:09
Repeating 'circuit board' then declaring 'I WANT TO GIVE UP' — peak AI malfunction captured.
Why this score: The dubbed speech devolves from attempting to convey a message into repetitive loops and what sounds like an AI expressing defeat, perfectly illustrating the technology's current limitations.
Original quote: “Hello, hello. I have a circuit board circuit board. Please leave it on your floor. I can't do this. I WANT TO GIVE UP.”
AI voice mangles simple names into gibberish — classic dubbing fail — Missing Context (45/100)
At 6:31
Shows ElevenLabs turning 'Cat Pat' into nonsense — highlights lip-sync and pronunciation issues.
Why this score: The clip demonstrates the core complaint: ElevenLabs' dubbing produces names that sound completely wrong. Without showing what the original name should have been or how a human dub would handle it, we're seeing the symptom but missing the full diagnosis.
Original quote: “I am Cat Pat. A few Bendy in the opposite direction. That's a b i e i, dear friend. Goodbye.”
Voice admits confusion mid-sentence — AI losing track of script — Confidence Mismatch (30/100)
At 6:43
The dubbed voice literally says 'I don't know what I'm saying' — peak self-own for AI dubbing.
Why this score: This is the perfect encapsulation of the video's thesis. The AI is so lost it starts narrating its own confusion. Without context on whether this is a translation error or voice model breakdown, it lands as pure comedy gold for the anti-ElevenLabs argument.
Original quote: “I don't know what I'm saying. Old and pure. Bendy. Thank you for watching.”
Nonsensical translation turns normal dialogue into word salad — Missing Context (35/100)
At 6:57
Original meaning completely lost — 'I receive sun lining' isn't even close to any sensible phrase.
Why this score: This shows translation failure rather than just voice quality issues. The viewer has no way to know if the source material was complex, if ElevenLabs' translation model failed, or if the dubbing process corrupted the output.
Original quote: “you I receive musical I receive sun lining small jump. Hello, your neighbor.”
AI invents profanity where none existed — translation gone rogue — Loaded Language (25/100)
At 7:13
Turns innocent dialogue into swearing — ElevenLabs adding content that wasn't there.
Why this score: This isn't just bad dubbing; it's actively changing the meaning and tone of the source material. The AI is inserting profanity that could completely alter how viewers perceive the original speaker's character.
Original quote: “a real dog, not at all the simplest. That person is more like a son of a [ __ ]”
AI hallucinates job titles and random animal references — Missing Context (30/100)
At 8:21
'Translation English supervisor' and 'toilet leopard' — pure AI word salad with zero connection to reality.
Why this score: This shows the compounding problem: bad translation feeding bad voice synthesis. The original likely had a completely different meaning, but without access to the source audio or script, we can't know how far off the rails ElevenLabs went.
Original quote: “Translation English supervisor is also just a damn toilet leopard gone missing.”
Shows AI switching between languages, technical glitches, and random screaming — complete system breakdown — Confidence Mismatch (10/100)
At 8:30
Demonstrates cascading failures: language switching, error messages, audio artifacts, and random vocalizations — shows system completely losing coherence.
Why this score: This segment captures multiple simultaneous failure modes: the AI begins mixing languages ('Loro waits'), produces technical error messages ('X-CODE HAS CALCULATED'), includes audio artifacts (throat clearing, screaming), and generates completely random content. This represents a total system collapse rather than simple translation errors.
Original quote: “Loro waits No. You know, one suitable fish on recipes salmon. AND THE >> [clears throat] >> X-CODE HAS CALCULATED >> a month >> [screaming]”
Generates increasingly bizarre religious/apocalyptic content completely detached from any source material — Missing Context (20/100)
At 9:20
AI produces elaborate metaphorical content about angels, hatred, and sunsets — complete invention with zero connection to original meaning.
Why this score: The dubbing has moved far beyond awkward phrasing into generating entirely new, elaborate content that has nothing to do with whatever the original audio contained. The AI is essentially creative writing rather than translating, producing content that would fundamentally change the meaning and tone of any source material.
Original quote: “It is a mange. With the sound of a phase in the chest? Martin, your knowledge is needed to join the parade with the angels bearing down on my hatred, the track to your sunset.”
Shows AI requesting human takeover while generating increasingly random content about robots and apples — Missing Context (25/100)
At 9:49
AI appears to request human intervention ('Reggie, can you take over?') while producing random content about robots and fruit — system acknowledging its own failure.
Why this score: The AI generates what sounds like a distress call or request for human assistance ('Reggie, can you take over?') while simultaneously producing increasingly random content. This creates an almost meta-commentary where the dubbing system appears to be asking for help while continuing to malfunction.
Original quote: “I'm up there. The sound changes the robot. Reggie, can you take over? Robot sound, you do work, okay? Robot sound. Loud explosion, let's talk. Taking care of apples?”
AI generates defensive response about 'accusations' — completely inventing context about being accused of something — Missing Context (20/100)
At 10:24
AI invents scenario where it's responding to 'accusations' — fabricating entire interpersonal conflict that doesn't exist in source material.
Why this score: The final example shows the AI generating what sounds like a defensive response to accusations that were never made. This demonstrates the system not only failing at translation but actively creating fictional interpersonal drama and conflict that has no basis in the original audio content.
Original quote: “But I live far away. I just WANT TO RESPOND TO THESE ACCUSATIONS. I”
AI garbles "I killed baby" into pet abuse context — tech failure — Missing Context (20/100)
At 10:30
The AI completely botches the line, turning it into something that sounds like animal harm — zero lip-sync or context preserved.
Why this score: The dubbing system inserts "I killed baby" right after "Good night" with no surrounding context, making it sound like a confession of pet violence. This is a classic case of the AI missing tone, timing, and narrative framing entirely.
Original quote: “I absolutely do not harm my pets, okay? Don't stop moving. Don't let me shock you again, please. Stop. Good night. I killed baby.”
AI spits out random words and phrases — complete semantic collapse — Missing Context (15/100)
At 11:04
The output becomes a word salad with no connection to the original meaning or context — pure gibberish.
Why this score: The AI has abandoned any attempt at coherent translation and is just stringing together unrelated words and brand names. This represents a total breakdown of the dubbing system's ability to maintain semantic meaning across languages.
Original quote: “Oh [__] Pac-Man is here. You, you, you, you, you. What is this, you [__] What school is it? SELFISH. 2 LITERS. I hate Jenny time I want grab dairy. English.”
AI randomly inserts moral commentary about drugs — context-free judgment — Missing Context (25/100)
At 11:25
The system adds a random anti-drug message that has nothing to do with the original content or context.
Why this score: The AI inserts an unrelated moral statement about drugs with zero connection to what's being discussed. This shows the dubbing system making editorial decisions without any understanding of the source material or intended meaning.
Original quote: “Indeed, Pac-Man. Drugs are bad. No, I'm unable to assist you, my friend. He says.”
AI dub mangles simple threat into nonsense — No Frame (75/100)
At 12:43
Straight demo of the tech failing — no spin, just showing the garbage output.
Why this score: The speaker isn't framing anything here; they're simply playing the broken ElevenLabs dub so the audience hears how it turns a straightforward line into word salad. It's raw evidence, not rhetoric.
Original quote: “I'm not just saying that I'm not feeling angry if you were to smash me, you would be able to scare me.”
Random baby crib line drops in — No Frame (75/100)
At 13:00
Still no framing — just more proof the dub is broken.
Why this score: The speaker continues the straight compilation format. No rhetorical devices, just accumulating examples.
Original quote: “I want to go here. That will be easy will be easy second, baby crib. Don't tell.”
Fragmented sentence that goes nowhere — No Frame (75/100)
At 13:08
Pure demo of the failure — no spin added by the speaker.
Why this score: The clip is presented as-is. The audience is meant to hear the mangled phrasing without any interpretive framing.
Original quote: “If I don't see it naturally if I won't Let's continue.”
Completely unrelated place and food names — No Frame (75/100)
At 13:20
No framing — speaker just lets the nonsense play out.
Why this score: Straightforward evidence presentation. The point is the dub's incoherence, not any rhetorical angle.
Original quote: “Okay Pittsburgh. Grub steak.”
AI mixes cooking instructions with random weather forecast — total nonsense — Missing Context (20/100)
At 14:37
Dubbing engine mashed unrelated sentences together, no attempt to preserve meaning.
Why this score: The AI was handed a coherent cooking segment and spat out weather gibberish. Shows the model doesn't track context across phrases.
Original quote: “I'm making this infirmary. Add three cups of cornstarch to the bowl and heat it. The weather today will be sunny on the side.”
MrBeast video becomes 'night wishant' — phonetic hallucination — Missing Context (20/100)
At 14:43
Proper noun mangled into nonsense words — classic ElevenLabs failure mode.
Why this score: The engine couldn't map the sounds to any known phrase, so it invented 'night wishant.' Highlights how fragile the model is with unusual names.
Original quote: “There's a heart-sick video of Mr. Beast that is almost ready to be released on COT. Oh my printer, night wishant, night wishant, must sing a normal song.”
Tone completely flips mid-sentence — emotional whiplash — Missing Context (20/100)
At 14:54
Victory line suddenly becomes profanity-laced taunt — no narrative bridge.
Why this score: The AI stitched two unrelated audio clips together without preserving speaker intent or emotional arc.
Original quote: “Yes, [music] I have done it, but what battle must I fight anymore? What the [ __ ] I look like ho ho ho.”
Random philosophical question followed by abrupt sign-off — no continuity — Missing Context (20/100)
At 15:32
Topic jumps from medicine to goodbye to ocean without transition.
Why this score: The model lost the thread entirely; each new phrase appears to come from a different source video.
Original quote: “How do medical conditions see the world? Thank you for watching. Do you have it? He is correct when he is in the open sea.”
Monkeys → mentorship → apocalypse in four seconds flat — Missing Context (20/100)
At 15:49
Rapid-fire topic roulette — model clearly can't maintain scene coherence.
Why this score: Each short burst sounds like it's from a different dubbed clip; context reset every couple words.
Original quote: “Uh what? monkeys Okay, see you in the next You will be my student from now on. The planet is finished.”
Existential despair instantly becomes ego trip — emotional 180 — Missing Context (20/100)
At 16:00
No logical bridge between 'no solution' and 'that's definitely me.'
Why this score: The AI glued two unrelated emotional beats together, creating a bizarre character arc in one breath.
Original quote: “I feel like there wasn't a solution. Oh, oh my goodness, that's me. I'm absolutely certain that's me. They're talking about me.”
The video presents a series of poorly dubbed clips, showcasing Elevenlabs' deficiencies. — No Frame (75/100)
At 16:30
The video directly shows examples of dubbing, letting the viewer judge the quality.
Why this score: This segment is a direct demonstration. The speaker's intent is to show, not tell, the issues with Elevenlabs' dubbing. The examples themselves are the evidence, and the framing is straightforward: 'here are the results, you decide'.
Original quote: “Subscribe to the Game Theory channel. >> me, spoke about me, and then lined me up. >> [music] >> Keep away from me. What? I am awake. What was that shouting >> [music] >> noise?”
The speaker uses strong emotional language to describe the dubbing quality. — Emotional Button (45/100)
At 17:03
Using words like 'ruined' is a classic emotional appeal, not a neutral assessment of quality.
Why this score: Calling the dubbing 'ruined' is a subjective, emotionally charged statement. While the examples might indeed be poor, this phrasing pushes the viewer towards a strong negative reaction rather than allowing for a more objective evaluation. A neutral framing would describe the specific technical flaws without such loaded language.
Original quote: “What is not? The screaming has ruined it. They have ruined it.”
The video showcases a highly offensive and hateful dubbed segment. — No Frame (75/100)
At 17:58
This segment directly demonstrates a problematic output, serving as evidence for the video's thesis.
Why this score: The inclusion of this highly offensive dubbed segment is a direct, unfiltered example of a potential failure or misuse of the dubbing technology. It's presented as raw evidence, allowing the viewer to witness the output directly. The framing here is simply presentation, letting the content speak for itself regarding the technology's limitations or vulnerabilities.
Original quote: “Why don't you adopt? I'm damn well going to say it. Why didn't you answer me, you game? Since you are black. YOU'RE ON, B, you fools. Can't you even see that it's flashing? YOU'RE SO INCREDIBLY DUMB, ABSOLUTELY WORTHLESS, you trashy, useless people.”
The AI explicitly states it cannot fulfill the request due to offensive content, immediately after a highly offensive dub. — Confidence Mismatch (20/100)
At 18:14
The AI's stated inability to process hate speech directly contradicts the preceding offensive output it just 'produced'.
Why this score: This is a classic 'confidence mismatch' because the AI's verbal statement about its safety guidelines (prohibiting hate speech) directly follows an example where it clearly *did* process and output highly offensive content. This creates a jarring contradiction, undermining the AI's stated capabilities or the effectiveness of its filters. The framing highlights a significant discrepancy between what the technology claims to do and what it demonstrably did in the preceding clip.
Original quote: “I cannot fulfill this request. The source text contains highly offensive and hateful language, including racial slurs. Generating or rephrasing such content goes against my safety guidelines, which prohibit the creation of hate speech.”
ElevenLabs refuses to dub content — safety filter triggered — No Frame (75/100)
At 18:41
Direct quote from the AI output itself — shows ElevenLabs' built-in content guardrails.
Why this score: The system is explicitly stating it won't translate because of offensive language. This isn't a technical failure; it's the model enforcing its safety policy. Clean framing because it's just reporting what happened without exaggeration.
Original quote: “THIS CONTENT contains offensive language [music] and cannot be translated.”
AI-generated line that doesn't match any original audio — nonsensical output — No Frame (75/100)
At 19:13
The dubbed audio produces completely unrelated, absurd dialogue with no connection to the source.
Why this score: This is a pure hallucination — the AI invented a line that has zero relationship to whatever was actually being said. The claim here is implicit: ElevenLabs sometimes generates text that makes no sense in context. The evidence is the line itself.
Original quote: “It smells terribly like a raccoon died in a diaper factory.”
Another nonsensical dubbed line — violent, bizarre phrasing — No Frame (75/100)
At 19:24
The AI is generating dialogue that escalates to cartoonish violence with no grounding in the original.
Why this score: This continues the pattern: ElevenLabs isn't just getting words wrong — it's creating entirely new, often disturbing content. The framing is straightforward because the clip speaks for itself as an example of the technology's failure mode.
Original quote: “Dude, I'm getting so incredibly angry right now. I could literally [ __ ] a squealing worm.”
Sudden shift from normal speech to repetitive 'boo boo boo' — audio loop failure — No Frame (75/100)
At 19:33
The dubbed voice goes from coherent to stuck in a loop of nonsense sounds.
Why this score: This is a classic failure mode where the model starts repeating filler or breaks rhythm entirely. The segment shows the technology collapsing mid-sentence without any external prompting.
Original quote: “Oh my goodness, I must say that looks absolutely fantastic. Boo boo boo boo boo boo boo boo boo”
AI demands dog food — completely detached from context — No Frame (75/100)
At 19:40
The dubbed character is now begging for pet food with no connection to the source material.
Why this score: Another clear hallucination. The AI has invented a new character motivation that has nothing to do with the original video. This is less a translation error and more the model going completely off-script.
Original quote: “I really want some Alpo, sir. I want some Alpo.”
See the full analysis with timestamps →