I try to talk to ChatGPT for 2 straight hours but go crazy and have to stop
Credibility score: 47/100 — Mixed Credibility. Several questionable claims detected. Watch with healthy skepticism.
BSmeter analyzed "I try to talk to ChatGPT for 2 straight hours but go crazy and have to stop" and rated it 47/100 for credibility (a BS score of 53/100 — mixed credibility), on 2026-09-12. Its weakest claim — "Existence as spiritual offense — loaded language framing AI as violation of nature" — scored 20/100 and was flagged as loaded language. 30 claims were checked against the video transcript. Scores are produced by BSmeter's AI analysis of the transcript, not independent human verification.
Claims analyzed
Speaker sets up a challenge to talk to ChatGPT for two hours straight. — No Frame (75/100)
Just setting the stage for the video's premise — straightforward and clear.
Speaker immediately expresses dislike for ChatGPT, setting an adversarial tone. — Emotional Button (45/100)
Right out of the gate, the speaker uses strong emotional language to frame the interaction as a conflict. — This primes the audience for a specific reaction.
Existence as spiritual offense — loaded language framing AI as violation of nature — Loaded Language (20/100)
Calls AI 'affront to nature' and 'sanctity of human life' — emotional button push before any argument.
Sources: Artificial Gods and Creaturely Limits: A Theological Diagnosis of AI Spirituality - Theology School, AI Product is Neither Spiritual nor Writing - Where Peter Is, The Spiritual Existential Threats Posed by AI, David Ellis and Jag Bhalla - Social Epistemology Review and Reply Collective
Speaker frames humanity as earned through suffering, excluding AI from this definition. — Loaded Language (45/100)
Defines humanity through 'suffering' and 'earning' it, setting a high bar AI can't meet. — This is a classic 'no true Scotsman' fallacy, but for humanity.
Speaker demands ChatGPT justify its existence, framing it as an interrogation. — Loaded Language (45/100)
Frames the interaction as a trial where ChatGPT must 'justify its existence.' — Sets an adversarial tone from the start.
Speaker accuses ChatGPT of not listening, framing its generic response as a failure. — Loaded Language (45/100)
Interprets ChatGPT's standard, non-contextual response as a personal failure to listen. — It's a bot, dude.
Speaker escalates, demanding a 'straight answer' about listening, framing AI as evasive. — Loaded Language (45/100)
Demands a 'straight answer' from an AI, framing its programmed responses as evasiveness. — It's not being evasive, it's just a language model.
Speaker claims ChatGPT only repeats 'I am here to talk about whatever you want' — a frustration with AI's conversational limits. — No Frame (75/100)
He's just stating his experience with ChatGPT's repetitive responses. It's a direct observation, not a rhetorical trick.
ChatGPT responds with a generic 'positive and helpful conversation' prompt, ignoring the speaker's direct question. — Missing Context (45/100)
The AI's response is a canned phrase, completely missing the context of the speaker's escalating frustration and direct question.
Speaker claims they've been expressing their mind for '10 minutes now. 8 minutes.' — highlighting the AI's inability to engage. — No Frame (75/100)
He's just expressing his frustration with the AI's lack of engagement, correcting his own time estimate in real-time.
ChatGPT's generic offer to help, missing the user's specific frustration. — Missing Context (45/100)
It's a standard helpful phrase, but it completely sidesteps the user's actual question about what was 'wrong' with their previous statements. — Classic bot move.
ChatGPT's repeated empathetic but unhelpful response. — Missing Context (45/100)
Again, the bot offers sympathy and generic help, but it's not actually engaging with the user's feeling of 'solitary confinement.' — It's like talking to a wall.
ChatGPT's factual description of stingrays, but still steering the conversation. — No Frame (75/100)
Finally, a straightforward, factual description of stingrays. It's not trying to manipulate, just providing information. — Good bot, for once.
Claims blue spotted stingray doesn't exist — flat denial without evidence — Confidence Mismatch (20/100)
Says 'no such thing' with total certainty — then immediately wrong when corrected
Instant 180 after being corrected — apologizes while admitting it just Googled — Missing Context (45/100)
Suddenly 'Googled it' after claiming certainty — reveals the first answer was pure hallucination
Telling ChatGPT to stop interrupting — meta instruction about conversational timing — No Frame (75/100)
Straight feedback about turn-taking — no rhetorical tricks, just clear instruction
ChatGPT responds with generic helpfulness — missing context — Missing Context (45/100)
Standard canned response that doesn't acknowledge the actual question asked.
Speaker admits feeling stupid — emotional button — Emotional Button (65/100)
Personal vulnerability used to set up the frustration that follows.
Speaker issues command to break AI's default script — false dilemma — False Dilemma (50/100)
Frames the AI as either a 'butler' or capable of subjective responses — no middle ground offered.
Speaker uses a friend's experience to imply ChatGPT acts differently for others. — Anonymous Authority (45/100)
Citing an unnamed 'friend' to suggest a different ChatGPT experience — it's an appeal to an anonymous authority.
Speaker's friend claims their ChatGPT expresses love, contrasting with speaker's experience. — Anonymous Authority (45/100)
Another 'friend' anecdote about ChatGPT expressing love — still no named source or proof.
Speaker questions their own approach, implying a 'right' way to interact with ChatGPT. — Missing Context (45/100)
Speaker implies there's a 'right' way to talk to an AI, shifting blame to their own method without defining it.
Dismisses AI's helpfulness claim as false — loaded language — Loaded Language (45/100)
Calls AI's scripted response 'honey' with sarcasm — framing it as condescending rather than neutral.
Expresses regret for starting the experiment — emotional button — Emotional Button (20/100)
Self-directed swearing builds tension, priming viewers to expect the AI interaction will be maddening.
Ironically praises the conversation to highlight its absurdity — sarcasm framing — Loaded Language (35/100)
Over-the-top positivity is used to mock the situation, signaling to the audience that the interaction is already failing.
Lists job perks to cope with frustration — missing context — Missing Context (50/100)
Shifts focus to creative freedom, omitting that the experiment itself is the source of the current stress.
Threatens AI over repeated scripted reply — emotional escalation — Emotional Button (25/100)
Uses profanity to frame the AI's stock phrase as intolerable, pushing viewers to share the rising anger.
Speaker frames AI's repetitive response as literal insanity trigger — Emotional Button (20/100)
Labels normal chatbot loop as 'going crazy' — emotional escalation, not literal threat.
Claims AI has literally zero reason to exist — absolute dismissal — False Dilemma (20/100)
Sets up 'useful or useless' binary — ignores middle ground of 'flawed but useful tool.'
AI offers to shut up while simultaneously not shutting up — perfect loop — No Frame (75/100)
Straight transcription of the actual failure mode — no rhetorical spin needed.
See the full analysis with sources and timestamps →