Youtubes Ai Thumbnails Get Even Worse
Credibility score: 41/100 — Mixed Credibility. Several questionable claims detected. Watch with healthy skepticism.
BSmeter analyzed "Youtubes Ai Thumbnails Get Even Worse" and rated it 41/100 for credibility (a BS score of 59/100 — mixed credibility), on 2026-07-20. Its weakest claim — "AI thumbnails "somehow" destabilizing global economy — emotional button + overreach" — scored 25/100 and was flagged as emotional button. 11 claims were checked against the video transcript. Scores are produced by BSmeter's AI analysis of the transcript, not independent human verification.
Of 11 claims analyzed: 5 scored under 40, 5 between 40 and 69, and 1 at 70 or above.
Claims analyzed
Calls his own video 'dumb' then sets up the AI thumbnail experiment — No Frame (75/100)
At 0:00
Self-deprecating opener that frames the whole bit as a silly experiment — classic creator humility move.
Why this score: He's not presenting this as serious journalism or a takedown. By calling his own video 'dumb' and labeling the next move a 'bright idea,' he signals this is playful content rather than a formal critique. A neutral version would just say 'I tested YouTube's AI thumbnails again' without the self-own framing.
Original quote: “So, a couple of weeks ago, I made a very dumb video where YouTube was suggesting me AI thumbnails, and they were the worst [__] I have seen in a while. And so, the only logical thing to do, of course, was to make a video out of how bad those suggestions were. Then, I had the bright idea of what if…”
Mocks YouTube's claim that 'It's over' is his famous catchphrase — it's not, and the AI invented the importance — Missing Context (45/100)
At 1:51
YouTube's justification inflates a minor recurring line into a 'signature' brand asset — classic overstatement.
Why this score: The rhetorical trick is elevation: a phrase he says sometimes gets reframed as the core of his identity so the thumbnail feels more 'on-brand.' He's right to flag it — the AI is manufacturing significance that doesn't exist at that level. Missing the actual frequency and weight of the phrase in his content.
Original quote: “The use of your signature 'It's over' text in the bottom center title case provides an instant brand recognition for your audience. You know the famous super eye patch will phrase, 'It's over.'”
Says YouTube is 'criticizing itself' — framing auto-generated text as self-aware critique — Missing Context (30/100)
At 2:55
Reads 'not me' as YouTube calling its own AI 'identity warping bollocks' — that's a huge interpretive leap.
Why this score: The auto-generated caption is just trying to describe the thumbnail. Treating it as intentional self-critique skips the simpler explanation: the AI prompt engine is bad at context. Missing that the system doesn't actually 'know' what it's saying.
Original quote: “YouTube says this layout directly contrasts your authentic self with an uncanny ghostly AI version... The not me text acts as a defiant brand statement signaling to viewers that this video is a deep dive into the identity warping nature of these AI tools. YouTube is criticizing itself.”
YouTube AI 'nailed it' on horror specter thumbnail — sarcastic praise — Emotional Button (45/100)
At 4:34
Calls the AI output 'nailed it' while describing it as literally horrifying — ironic framing to mock the result.
Why this score: He's setting up the thumbnail as both technically successful and emotionally disturbing at the same time. The word 'nailed it' is doing the heavy lifting — it's the same phrase creators use for genuinely good thumbnails, so flipping it here makes the AI look even more unhinged by comparison. Neutral version would just say 'this one is bad' without the fake praise.
Original quote: “This thumbnail captures your most extreme reaction of horror from the video paired with the looming hellish AI-generated fire ghost version of yourself that you described as a specter of death. You know what? I'm going to say once again, YouTube, you nailed it.”
Calls thumbnail 'nightmarish' and 'disturbing' — emotional escalation — Emotional Button (40/100)
At 6:56
Stacks words like 'nightmarish' and 'spookered' — turns a bad thumbnail into something actively traumatizing
Why this score: The speaker escalates from 'weird' to 'nightmarish' in seconds, building a horror-movie reaction instead of describing what the image actually contains. This framing makes the thumbnail feel more extreme than it might to an unprimed viewer.
Original quote: “Yeah, I know. I don't know why I'm have to [ __ ] give that either. Okay, shit's getting really weird and disturbing now. I feel myself properly spookered, and I just hope YouTube's AI doesn't have any more nightmarish images to harv- Oh my God, no. No.”
Interprets distorted analytics graph as literal 'it's over' — reads doom into glitch — Missing Context (50/100)
At 7:18
Treats AI rendering error as meaningful commentary on his channel's performance — the graph is just broken, not prophetic
Why this score: The speaker presents the warped analytics display as evidence that 'it's really over,' when it's actually just the AI glitching while trying to render text and charts. A neutral framing would note that the image is nonsensical rather than treating the distortion as symbolic.
Original quote: “For one, I don't know what's happening. This AI version of me here, I appear to have a snake coffin for a brain. And also, just looking at this here, it seems I'm looking at my own analytics page, and they're they're not doing great. In fact, my analytics are so bad that they're skewering the…”
Frames thumbnail as 'void of imagery where humanity does not exist' — dramatic literary comparison — Emotional Button (35/100)
At 8:12
Compares bad AI art to existential horror and Hunter x Hunter curse — turns thumbnail fail into cosmic dread
Why this score: By invoking 'the void of imagery where humanity does not exist' and a cursed object from anime, the speaker elevates a glitchy thumbnail into something philosophically terrifying. A cleaner frame would keep the reaction at 'this looks messed up' instead of reaching for cosmic nihilism.
Original quote: “This is this is just what happens when you reach into the realm beyond humanity and pull back an object of nihilism. This is like this is like a cursed object from the dark continent in Hunter Hunter. This is the void of imagery where humanity does not exist and I don't like it.”
Describes AI image as "transgender skeletons" dragging him — loaded language — Loaded Language (30/100)
At 8:46
Uses "transgender skeletons" as the core descriptor — instantly frames the weirdness around identity rather than the actual absurdity.
Why this score: The speaker chooses a highly specific, charged label for the skeletons instead of a neutral description like "human-like figures" or "stylized characters." This single word choice steers the reaction toward discomfort before any other details are processed. A neutral framing would simply describe the visual elements without injecting identity politics into the critique of AI imagery.
Original quote: “My God. So, this one appears to be a group of transgender skeletons dragging me into a snake coffin as I make the most AI rage face I have ever seen.”
Claims the image fails to deliver "metal" energy — emotional button — Emotional Button (40/100)
At 9:47
Sets up an expectation of "metal" then knocks it down to "weird and dumb" — uses emotional contrast to amplify disappointment.
Why this score: The speaker first builds a hypothetical positive reaction ("sounds metal") then immediately undercuts it with a visceral negative judgment. This emotional whiplash makes the AI image feel more disappointing than a straightforward description would. A cleaner framing would just say the image doesn't work without the metal/dumb binary.
Original quote: “I'd be like, "Man, that sounds pretty [__] metal." And then I look at this thing and I'm like, "No, that's not metal. That's just [__] weird and dumb."”
Contrasts human art (happy) with AI (uncanny valley mediocrity) — false equivalence — False Equivalence (35/100)
At 10:07
Treats one human redraw as proof that all AI feels "mediocre uncanny valley" — equates one example with the entire category.
Why this score: The speaker presents Anna Montana's redraw as the emotional opposite of every AI image shown, implying all AI output shares the same flaw. This skips over the fact that AI quality varies widely and that the human version was deliberately made to fix the original. A fairer framing would compare the specific flaws of this AI thumbnail to what a human artist typically improves, without generalizing to "every other" AI image.
Original quote: “This image makes me feel happy. This image makes me feel like a real human did something and communicated that something through art. ...which is completely different to the weird level of mediocre uncanny valley [__] I have felt with every other”
AI thumbnails "somehow" destabilizing global economy — emotional button + overreach — Emotional Button (25/100)
At 10:46
Calls thumbnails "dumb" then pivots to "destabilizing our global economy" — same sentence, zero bridge.
Why this score: The speaker moves from a mild aesthetic complaint straight to a massive economic claim without any mechanism, data, or even a connecting phrase. By using "somehow," he signals he knows the leap is big but still wants the emotional weight of "global economy" to land. A neutral framing would separate the personal taste judgment from the economic claim entirely.
Original quote: “Okay, that's it. That's the last time I talk about these stupid AI thumbnails cuz I think they're weird and gross. It really makes me feel some kind of way in that yeah, these things are dumb, these things are kind of pointless, but this is also somehow the technology destabilizing our global…”
See the full analysis with timestamps →