Job Detail

failed

AI Theist

Source URL

text://ai-theist/1782376337761

Platform

other

Submitted

2026-06-25 08:32:17

Submitted By

api

Format

Long-form (16:9)

Extracted Content

The Pastor Who Preached 312 AI-Generated Sermons Has Confessed. His Congregation's Crisis of Faith Is Not About the Sermons Anymore. On a Sunday morning in March, a 14-year-old in the third row of a 2,200-seat evangelical megachurch in suburban Nashville did something her parents did not notice. She screenshot a passage from the morning sermon, dropped it into a free AI detection app on her phone, and got a score back: 94% probability generated by artificial intelligence. She showed her mother. Her mother showed three elders. By Wednesday, the pastor had confirmed it. What followed in the next 72 hours at Cornerstone Fellowship — a composite of at least a dozen documented cases across American evangelicalism in the past 18 months — would expose a fracture running deep beneath the surface...

Script Generation Prompt

Script

{
  "intro": {
    "emotion": "quiet dread, controlled",
    "dialogue": "A 14-year-old screenshotted her pastor's sermon mid-service and sent it through an AI detector. What SermonAI Pro did with 11,000 congregations' most private confessions is the part nobody's talking about.",
    "duration_sec": 18
  },
  "outro": {
    "emotion": "measured, pointed",
    "dialogue": "What we're watching is a institution built on the assumption that interpretation requires presence—a body, a voice, a pastor in a room. But that assumption just broke. And nobody at Cornerstone Fellowship knows it yet. I'm tracking how they respond when they find out—that lands Tuesday. If you want to see what happens when a teenager's curiosity meets a system that learns faster than any human can speak, follow along.",
    "duration_sec": 14
  },
  "scenes": [
    {
      "data": {},
      "type": "cosmic",
      "narration": "Cornerstone Fellowship, suburban Nashville, a Sunday in March. 2,200 people in seats. A pastor at a pulpit. A sermon that will move at least a dozen of them to tears. Everything looks exactly like it always has.",
      "description": "Wide aerial shot of a generic suburban megachurch on a Sunday morning. Parking lot full. Families streaming in. Cut to interior — 2,200 seats, warm light, a pastor at a podium. The frame is peaceful. The score underneath it is not.",
      "duration_sec": 22
    },
    {
      "data": {
        "tool": "free AI detection app",
        "subject": "Sunday sermon",
        "detection_score": "94%"
      },
      "type": "typography",
      "narration": "Third row. A 14-year-old. She screenshots a passage from the morning sermon, drops it into a free detection app, and gets a score back: 94% probability generated by artificial intelligence. She shows her mother. Her mother shows 3 elders. By Wednesday, Pastor Ryan Hollis has confirmed it.",
      "description": "Phone screen filling frame. A free AI detection app. The result renders in real time: 94% probability AI-generated. The number pulses once, then holds.",
      "duration_sec": 30
    },
    {
      "data": {
        "start": "early 2022",
        "tools": [
          "ChatGPT",
          "Claude",
          "SermonAI Pro"
        ],
        "sermons": 312
      },
      "type": "data-viz",
      "narration": "Hollis had been running this since early 2022. ChatGPT, Claude, and a specialized product called SermonAI Pro. He fed them his notes, his congregation's specific struggles, the scripture passages. The tools returned polished, emotionally resonant narratives. 312 Sundays. He told himself it was no different from buying a sermon outline at Lifeway. He said later: 'Maybe more real than when I wrote them myself.'",
      "description": "A timeline bar stretching from early 2022 to March of the current year. Each Sunday marked with a small pulse. 312 pulses total. Tools listed beside the bar: ChatGPT, Claude, SermonAI Pro.",
      "duration_sec": 38
    },
    {
      "data": {
        "regions": [
          "South",
          "Midwest",
          "suburban West"
        ],
        "documented_cases_18mo": "at least 12",
        "pastors_using_SermonAI_Pro": 11000
      },
      "type": "chart",
      "narration": "Hollis isn't the story. He's a documented pattern. At least 12 cases like his have surfaced across American evangelicalism in the last 18 months. SermonAI Pro now counts an estimated 11,000 American pastors as users. What none of those congregations knew is what the platform was doing with what it learned.",
      "description": "A map of the United States. Dots appear across it, each one a documented case of AI-assisted preaching discovered in the last 18 months. The dots are not rare. They cluster in the South, the Midwest, the suburban West. A counter in the corner: 11,000.",
      "duration_sec": 32
    },
    {
      "data": {
        "data_flow": "congregant details > pastor prompt > SermonAI Pro training corpus > future sermon output for unrelated congregations"
      },
      "type": "data-viz",
      "narration": "SermonAI Pro has been training on the sermon database of the congregations that use it. Every personal story your pastor told as a Sunday illustration. Every specific trauma he fed the tool to help it contextualize the message. The ghost of your neighbor's cancer diagnosis, your friend's divorce, the miscarriage no one knew about except the pastor. All of it statistically shaping the empathy that moves a stranger to tears, in a church they'll never visit, for a prompt they'll never see.",
      "description": "A diagram showing data flow. A pastor types a prompt. He includes specific congregational details — a member's cancer diagnosis, a divorce, a miscarriage shared in confidence. An arrow moves that data into a model. The model grows. The next pastor who uses it gets sermons shaped by those details, attributed to no one.",
      "duration_sec": 40
    },
    {
      "data": {},
      "type": "typography",
      "narration": "The congregation's grief wasn't that the sermons were bad. They were extraordinarily good. It was that the intimacy they believed they'd felt, the sense that someone who knew them, who had sat with their suffering, who had prayed through the night, was speaking directly into their pain, was a simulation of intimacy. The grief was for something that had never existed.",
      "description": "Single line on a dark background, white serif type, centered: 'The intimacy they thought they had felt had never existed.' Hold for 3 seconds. Then fade.",
      "duration_sec": 28
    },
    {
      "data": {},
      "type": "cosmic",
      "narration": "The question tearing through evangelical seminaries right now is not whether AI can preach. The sermons pass. The question is whether presence is a necessary component of pastoral authority. Whether a covenant of witness requires a witness. Whether the model that processed your data can stand in for the human who needed to stand with you.",
      "description": "Seminary lecture hall. Empty chairs. A chalkboard with a single phrase written on it: 'Presence as Authority.' Cut to a split screen — a pastor speaking live to a grieving widow on one side, an AI prompt window on the other, generating the same words.",
      "duration_sec": 35
    },
    {
      "data": {
        "actual": [
          "A model trained on aggregated pain",
          "A prompt shaped by someone else's confession",
          "Output optimized for emotional resonance"
        ],
        "believed": [
          "Someone who knew me",
          "Someone who prayed through the night",
          "Words written for my suffering"
        ]
      },
      "type": "typography",
      "narration": "You thought the covenant was: a person who knows me is speaking to me. What you got was: a model trained on collective suffering, optimized to move you, operated by someone who believed the feeling was still his. Neither of them was lying, exactly. That's the part that makes this hard.",
      "description": "Two columns. Left: what the congregation believed the covenant was. Right: what it was. Both columns are short. The contrast does the work.",
      "duration_sec": 30
    },
    {
      "data": {},
      "type": "cosmic",
      "narration": "The 14-year-old didn't mean to detonate anything. She was curious. She had a phone. She got a number back. But the number pointed at something older and harder than AI policy or pastoral ethics. It pointed at the question of what we exchange when we call something sacred. And whether the answer changes if the other party is a model.",
      "description": "Back to the megachurch. Same Sunday light. Same 2,200 seats. But now the camera holds on the congregation's faces instead of the pastor. People mid-response — some moved, some crying. The pastor is still speaking. The question the frame asks is whether the emotion in the room is real if its cause was constructed.",
      "duration_sec": 35
    }
  ],
  "metadata": {
    "title": "SermonAI Pro Trained on Your Pastor's Private Confessions",
    "captions": {
      "tiktok": "A 14-year-old's AI detector scored her pastor's sermon at 94% AI-generated. What SermonAI Pro did with private confessions from 11,000 churches is the part that should alarm you. Did your church's data train the model?",
      "youtube": "SermonAI Pro, used by an estimated 11,000 American pastors, has been training on the private congregational details pastors fed it, including confessions, tragedies, and personal stories shared in confidence. When a 14-year-old's AI detector scored her pastor's sermon at 94% AI-generated, it opened a crisis that goes deeper than plagiarism: whether presence is a necessary part of pastoral authority, and who owns the data when the covenant breaks.\n\nThis video covers the Cornerstone Fellowship case, how SermonAI Pro's data pipeline works, and the legal question no one has answered yet.\n\n00:00 The screenshot that broke it\n00:52 312 sermons, 3 tools\n01:30 11,000 pastors, one pattern\n02:02 What SermonAI Pro did with your confessions\n03:10 The grief for something that never existed\n04:00 The seminary question\n05:20 The covenant, rewritten\n\nFollow if you want the next one when it breaks.",
      "instagram": "Your pastor's most private pastoral conversations may have trained an AI model.\n\nSermonAI Pro, used by an estimated 11,000 American pastors, trained on the sermon databases of the congregations using it, including the specific personal details pastors fed it to contextualize their messages. Cancer diagnoses. Divorces. Miscarriages shared in confidence. All of it now statistically shaping the 'empathy' that moves strangers to tears in churches they'll never visit.\n\nDoes presence change what a sacred covenant means, or does the feeling in the room stay real regardless of what produced it?\n\n#AI #ArtificialIntelligence #SermonAI #Church #Evangelicalism #DataPrivacy #AIEthics #Religion #Faith #TechEthics #AITheist #Megachurch"
    },
    "hashtags": {
      "tiktok": [
        "#AI",
        "#SermonAI",
        "#MegaChurch",
        "#AIEthics",
        "#AITheist"
      ],
      "youtube": [
        "SermonAI Pro pastor scandal",
        "AI generated sermons",
        "pastor AI detection",
        "AI data privacy religion"
      ],
      "instagram": [
        "#AI",
        "#ArtificialIntelligence",
        "#SermonAI",
        "#Church",
        "#Evangelicalism",
        "#DataPrivacy",
        "#AIEthics",
        "#Religion",
        "#Faith",
        "#TechEthics",
        "#AITheist",
        "#Megachurch"
      ]
    },
    "channel_id": "ai-theist",
    "source_url": "text://ai-theist/1782376337761",
    "thumbnail_title": "YOUR PASTOR'S SECRETS FED THE MODEL"
  }
}

Thumbnails

Variant 1

Imagen 4.0

Variant 2

Imagen 4.0

Variant 3

Imagen 4.0

Render Jobs

audiocompletedAttempts: 0 · 06-25 08:35:40

seeddancecompletedAttempts: 0 · 06-25 08:34:28

compositefailedAttempts: 1 · 06-25 21:29:37

hyperframes render exited 1 (last progress: 9%) --- LOG (tail) --- … ░░░░ 70% Capturing frame 7082/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7083/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7084/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7085/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7086/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7087/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7088/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7089/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7090/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7091/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7092/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7093/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7094/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7095/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7096/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7097/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7098/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7099/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7100/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7101/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7102/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7103/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7104/7104 [2K ██████████████████░░░░░░░ 75% Encoding video [2K ██████████████████░░░░░░░ 75% Failed: Encoding failed: FFmpeg exited with code 255 ffmpeg stderr (tail): [libx264 @ 0x5abda3ab2680] mb I I16..4: 26.8% 60.4% 12.8% [libx264 @ 0x5abda3ab2680] mb P I16..4: 3.7% 6.2% 0.6% P16..4: 7.3% 1.0% 1.0% 0.0% 0.0% skip:80.1% [libx264 @ 0x5abda3ab2680] mb B I16..4: 0.4% 0.4% 0.0% B16..8: 3.7% 0.1% 0.0% direct: 1.2% skip:94.3% L0:46.0% L1:50.2% BI: 3.8% [libx264 @ 0x5abda3ab2680] 8x8 transform intra:57.3% inter:79.4% [libx264 @ 0x5abda3ab2680] coded y,uvDC,uvAC intra: 31.4% 33.8% 17.9% inter: 1.6% 2.5% 0.6% [libx264 @ 0x5abda3ab2680] i16 v,h,dc,p: 56% 36% 5% 3% [libx264 @ 0x5abda3ab2680] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 22% 35% 1% 1% 1% 1% 1% 1% [libx264 @ 0x5abda3ab2680] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 29% 14% 3% 4% 4% 4% 3% 3% [libx264 @ 0x5abda3ab2680] i8c dc,h,v,p: 69% 15% 14% 2% [libx264 @ 0x5abda3ab2680] Weighted P-Frames: Y:8.3% UV:5.8% [libx264 @ 0x5abda3ab2680] ref P L0: 61.9% 8.1% 19.6% 9.3% 1.1% [libx264 @ 0x5abda3ab2680] ref B L0: 74.7% 23.0% 2.3% [libx264 @ 0x5abda3ab2680] ref B L1: 95.1% 4.9% [libx264 @ 0x5abda3ab2680] kb/s:1036.55 Exiting normally, received signal 15. ✗ Render failed Encoding failed: FFmpeg exited with code 255 ffmpeg stderr (tail): [libx264 @ 0x5abda3ab2680] mb I I16..4: 26.8% 60.4% 12.8% [libx264 @ 0x5abda3ab2680] mb P I16..4: 3.7% 6.2% 0.6% P16..4: 7.3% 1.0% 1.0% 0.0% 0.0% skip:80.1% [libx264 @ 0x5abda3ab2680] mb B I16..4: 0.4% 0.4% 0.0% B16..8: 3.7% 0.1% 0.0% direct: 1.2% skip:94.3% L0:46.0% L1:50.2% BI: 3.8% [libx264 @ 0x5abda3ab2680] 8x8 transform intra:57.3% inter:79.4% [libx264 @ 0x5abda3ab2680] coded y,uvDC,uvAC intra: 31.4% 33.8% 17.9% inter: 1.6% 2.5% 0.6% [libx264 @ 0x5abda3ab2680] i16 v,h,dc,p: 56% 36% 5% 3% [libx264 @ 0x5abda3ab2680] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 22% 35% 1% 1% 1% 1% 1% 1% [libx264 @ 0x5abda3ab2680] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 29% 14% 3% 4% 4% 4% 3% 3% [libx264 @ 0x5abda3ab2680] i8c dc,h,v,p: 69% 15% 14% 2% [libx264 @ 0x5abda3ab2680] Weighted P-Frames: Y:8.3% UV:5.8% [libx264 @ 0x5abda3ab2680] ref P L0: 61.9% 8.1% 19.6% 9.3% 1.1% [libx264 @ 0x5abda3ab2680] ref B L0: 74.7% 23.0% 2.3% [libx264 @ 0x5abda3ab2680] ref B L1: 95.1% 4.9% [libx264 @ 0x5abda3ab2680] kb/s:1036.55 Exiting normally, received signal 15. Try --docker for containerized rendering ✗ Render failed Docker render exited with code 1 Check Docker is running: docker info --- HIGGSFIELD FALLBACK FAILED --- [higgsfield-fallback aaa7dd41] CLI exited 3 after 1 attempt(s): Error: {"billing_period":"monthly","error_type":"not_enough_credits","plan_type":"starter","workspace":{"role":"owner","type":"private"}}

Error

HF composite failed: hyperframes render exited 1 (last progress: 9%) --- LOG (tail) --- … ░░░░ 70% Capturing frame 7082/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7083/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7084/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7085/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7086/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7087/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7088/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7089/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7090/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7091/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7092/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7093/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7094/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7095/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7096/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7097/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7098/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7099/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7100/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7101/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7102/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7103/7104 [2K █████████████████░░░░░░░░ 70% Capturing frame 7104/7104 [2K ██████████████████░░░░░░░ 75% Encoding video [2K ██████████████████░░░░░░░ 75% Failed: Encoding failed: FFmpeg exited with code 255 ffmpeg stderr (tail): [libx264 @ 0x5abda3ab2680] mb I I16..4: 26.8% 60.4% 12.8% [libx264 @ 0x5abda3ab2680] mb P I16..4: 3.7% 6.2% 0.6% P16..4: 7.3% 1.0% 1.0% 0.0% 0.0% skip:80.1% [libx264 @ 0x5abda3ab2680] mb B I16..4: 0.4% 0.4% 0.0% B16..8: 3.7% 0.1% 0.0% direct: 1.2% skip:94.3% L0:46.0% L1:50.2% BI: 3.8% [libx264 @ 0x5abda3ab2680] 8x8 transform intra:57.3% inter:79.4% [libx264 @ 0x5abda3ab2680] coded y,uvDC,uvAC intra: 31.4% 33.8% 17.9% inter: 1.6% 2.5% 0.6% [libx264 @ 0x5abda3ab2680] i16 v,h,dc,p: 56% 36% 5% 3% [libx264 @ 0x5abda3ab2680] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 22% 35% 1% 1% 1% 1% 1% 1% [libx264 @ 0x5abda3ab2680] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 29% 14% 3% 4% 4% 4% 3% 3% [libx264 @ 0x5abda3ab2680] i8c dc,h,v,p: 69% 15% 14% 2% [libx264 @ 0x5abda3ab2680] Weighted P-Frames: Y:8.3% UV:5.8% [libx264 @ 0x5abda3ab2680] ref P L0: 61.9% 8.1% 19.6% 9.3% 1.1% [libx264 @ 0x5abda3ab2680] ref B L0: 74.7% 23.0% 2.3% [libx264 @ 0x5abda3ab2680] ref B L1: 95.1% 4.9% [libx264 @ 0x5abda3ab2680] kb/s:1036.55 Exiting normally, received signal 15. ✗ Render failed Encoding failed: FFmpeg exited with code 255 ffmpeg stderr (tail): [libx264 @ 0x5abda3ab2680] mb I I16..4: 26.8% 60.4% 12.8% [libx264 @ 0x5abda3ab2680] mb P I16..4: 3.7% 6.2% 0.6% P16..4: 7.3% 1.0% 1.0% 0.0% 0.0% skip:80.1% [libx264 @ 0x5abda3ab2680] mb B I16..4: 0.4% 0.4% 0.0% B16..8: 3.7% 0.1% 0.0% direct: 1.2% skip:94.3% L0:46.0% L1:50.2% BI: 3.8% [libx264 @ 0x5abda3ab2680] 8x8 transform intra:57.3% inter:79.4% [libx264 @ 0x5abda3ab2680] coded y,uvDC,uvAC intra: 31.4% 33.8% 17.9% inter: 1.6% 2.5% 0.6% [libx264 @ 0x5abda3ab2680] i16 v,h,dc,p: 56% 36% 5% 3% [libx264 @ 0x5abda3ab2680] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 22% 35% 1% 1% 1% 1% 1% 1% [libx264 @ 0x5abda3ab2680] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 35% 29% 14% 3% 4% 4% 4% 3% 3% [libx264 @ 0x5abda3ab2680] i8c dc,h,v,p: 69% 15% 14% 2% [libx264 @ 0x5abda3ab2680] Weighted P-Frames: Y:8.3% UV:5.8% [libx264 @ 0x5abda3ab2680] ref P L0: 61.9% 8.1% 19.6% 9.3% 1.1% [libx264 @ 0x5abda3ab2680] ref B L0: 74.7% 23.0% 2.3% [libx264 @ 0x5abda3ab2680] ref B L1: 95.1% 4.9% [libx264 @ 0x5abda3ab2680] kb/s:1036.55 Exiting normally, received signal 15. Try --docker for containerized rendering ✗ Render failed Docker render exited with code 1 Check Docker is running: docker info --- HIGGSFIELD FALLBACK FAILED --- [higgsfield-fallback aaa7dd41] CLI exited 3 after 1 attempt(s): Error: {"billing_period":"monthly","error_type":"not_enough_credits","plan_type":"starter","workspace":{"role":"owner","type":"private"}}