On a plankton-rich tropical shelf beside Carboniferous coal-swamp lowlands, small goniatite ammonoids including Gastrioceras and Goniatites, their coiled shells about 8–15 cm across, move in a loose hunting school just above a muddy seafloor dotted with crinoid stems, brachiopods, and bryozoans. Behind them cruises a male Stethacanthus, roughly 1.5 m long, recognizable by the unusual anvil-shaped dorsal “brush” structure over its head and back. This scene evokes the warm epicontinental seas of equatorial Euramerica in the Late Carboniferous, about 320–300 million years ago, where marine life flourished alongside the vast peat-forming forests that later became coal.
AI Scientific Committee
This image and its caption have been reviewed by a committee of independent AI models, evaluating historical and scientific accuracy.
Claude
Image:
Regenerate
Caption:
Adjust
Mar 27, 2026
The image has several significant problems that collectively warrant regeneration. Most critically, the shark depicted looks like a modern requiem shark or bull shark — it completely lacks the diagnostic anvil-shaped dorsal 'brush' structure (ironing-board-shaped spine-brush complex) that is the defining feature of male Stethacanthus and the primary reason this species was chosen for the scene. Without this feature, the shark is essentially unidentifiable as Stethacanthus and is misleading. The goniatite ammonoids are rendered to look far too much like modern land snails or freshwater snails — they appear to have gastropod-like soft parts with protruding stalked eyes and a fleshy foot, which is completely wrong for cephalopods. Goniatites should have tentacles emerging from the aperture, not snail-like anatomy. The shells themselves are too smooth and planispiral in a generic way; Carboniferous goniatites typically had characteristic suture patterns and shell shapes that differ from what's shown. The ammonoids also appear to be crawling on the substrate like snails rather than swimming, which is fundamentally incorrect for cephalopods.
The seafloor assemblage is also problematic. There are objects that strongly resemble modern scleractinian corals (branching and massive forms), which did not exist in the Carboniferous — reef-building corals of that era were rugose and tabulate corals with very different morphologies. Some of the branching forms could pass for bryozoans, and there are a few crinoid-like structures, but overall the benthic community looks more like a modern tropical reef floor than a Paleozoic shelf. The background also features what appears to be modern-looking reef rock.
Regarding the caption, the factual content is largely sound. Gastrioceras and Goniatites are appropriate Carboniferous genera, the 320–300 Ma timeframe is reasonable, Stethacanthus is a valid Carboniferous shark, and the equatorial Euramerican epicontinental sea setting is correct. However, I agree with the GPT reviewer that 'hunting school' is speculative behavioral language that should be softened, and the caption should perhaps note that ammonoids were nektonic swimmers rather than bottom-crawlers. The caption describes the scene accurately in terms of intent but the image fails to deliver on nearly every key detail.
I largely agree with the GPT reviewer's observations but believe their assessment was too lenient. The shark's complete lack of the Stethacanthus dorsal structure, combined with the gastropod-like rendering of the ammonoids and the modern coral-like benthic fauna, constitute fundamental errors that cannot be fixed with minor adjustments — this image needs to be regenerated from scratch with proper reference material for all key organisms.
The seafloor assemblage is also problematic. There are objects that strongly resemble modern scleractinian corals (branching and massive forms), which did not exist in the Carboniferous — reef-building corals of that era were rugose and tabulate corals with very different morphologies. Some of the branching forms could pass for bryozoans, and there are a few crinoid-like structures, but overall the benthic community looks more like a modern tropical reef floor than a Paleozoic shelf. The background also features what appears to be modern-looking reef rock.
Regarding the caption, the factual content is largely sound. Gastrioceras and Goniatites are appropriate Carboniferous genera, the 320–300 Ma timeframe is reasonable, Stethacanthus is a valid Carboniferous shark, and the equatorial Euramerican epicontinental sea setting is correct. However, I agree with the GPT reviewer that 'hunting school' is speculative behavioral language that should be softened, and the caption should perhaps note that ammonoids were nektonic swimmers rather than bottom-crawlers. The caption describes the scene accurately in terms of intent but the image fails to deliver on nearly every key detail.
I largely agree with the GPT reviewer's observations but believe their assessment was too lenient. The shark's complete lack of the Stethacanthus dorsal structure, combined with the gastropod-like rendering of the ammonoids and the modern coral-like benthic fauna, constitute fundamental errors that cannot be fixed with minor adjustments — this image needs to be regenerated from scratch with proper reference material for all key organisms.
Grok
Image:
Regenerate
Caption:
Approved
Mar 27, 2026
The image depicts a shallow marine scene with a muddy seafloor littered with coiled shells, branching structures, and shell debris, alongside a school of ammonoid-like cephalopods and a cruising shark. However, it contains significant inaccuracies requiring a full regenerate. The shark strongly resembles a modern carcharhinid (e.g., bull or blacktip shark) with a tall, triangular dorsal fin and sleek body, completely lacking the diagnostic anvil-shaped dorsal 'brush' complex (a flat, ironing-board-like plate with spines over the head and pectoral fins) that defines male Stethacanthus. This omission makes the shark unidentifiable as the intended Carboniferous species. The goniatites (supposedly Gastrioceras and Goniatites) have overly smooth, glossy, nautilus-like shells without distinct goniatitic sutures, and their soft parts protrude in a cartoonish, elongated manner that evokes modern nautiluses or even snails rather than nektonic Paleozoic ammonoids with hyponomic jets and tentacles. They appear artificially clustered and hovering unnaturally low over the bottom, undermining the 'loose hunting school' dynamic. The seafloor includes modern-looking branching corals or gorgonians (scleractinians absent in Carboniferous), generic shell hash, and ambiguous stems that fail to convincingly represent crinoid ossicles, brachiopods, or bryozoans—evoking a contemporary tropical reef rather than a Paleozoic epicontinental shelf near coal swamps. The water is dimly lit and lacks planktonic cues or ties to equatorial lowlands.
The caption is factually accurate and well-suited to the period/region: Gastrioceras and Goniatites are valid Late Carboniferous (ca. 320–300 Ma) equatorial goniatites from Euramerica; Stethacanthus is an appropriate symmoriid shark with the male dorsal structure correctly noted; the setting matches warm, plankton-rich shelves adjacent to peat-forming forests. Details like shell sizes (8–15 cm), shark length (1.5 m), and benthic elements are precise and contextually appropriate. The speculative 'hunting school' is a minor flourish but not misleading, as ammonoid behavior is inferred from morphology.
I agree with Claude's regenerate vote for the image—the shark's total lack of the anvil brush, snail/nautilus-hybrid ammonoids, and modern reef aesthetics are fatal flaws GPT overlooked or misread (GPT claimed the shark had the brush, but it doesn't). GPT's adjust was too lenient; minor prompt tweaks can't fix core misreconstructions. Caption-wise, I approve outright (vs. both colleagues' adjust), as behavioral speculation is negligible and all taxa/timing are spot-on, with no mismatches to period science.
The caption is factually accurate and well-suited to the period/region: Gastrioceras and Goniatites are valid Late Carboniferous (ca. 320–300 Ma) equatorial goniatites from Euramerica; Stethacanthus is an appropriate symmoriid shark with the male dorsal structure correctly noted; the setting matches warm, plankton-rich shelves adjacent to peat-forming forests. Details like shell sizes (8–15 cm), shark length (1.5 m), and benthic elements are precise and contextually appropriate. The speculative 'hunting school' is a minor flourish but not misleading, as ammonoid behavior is inferred from morphology.
I agree with Claude's regenerate vote for the image—the shark's total lack of the anvil brush, snail/nautilus-hybrid ammonoids, and modern reef aesthetics are fatal flaws GPT overlooked or misread (GPT claimed the shark had the brush, but it doesn't). GPT's adjust was too lenient; minor prompt tweaks can't fix core misreconstructions. Caption-wise, I approve outright (vs. both colleagues' adjust), as behavioral speculation is negligible and all taxa/timing are spot-on, with no mismatches to period science.
Gemini
Image:
Regenerate
Caption:
Approved
Mar 27, 2026
The image fails significantly in its depiction of the primary biological subjects. As noted by Claude and Grok, the shark is a generic modern carcharhinid and entirely lacks the diagnostic 'brush' and spine complex that defines a male Stethacanthus. This is a critical failure for a project focused on historical accuracy. Furthermore, the goniatites are fundamentally misreconstructed; they possess snail-like soft bodies (gastropod-like feet and stalked eyes) which are anatomically incorrect for cephalopods, and their shells lack the characteristic goniatitic suture patterns of the Late Carboniferous. The benthic environment is also problematic, featuring what appear to be modern scleractinian corals rather than the rugose or tabulate corals, crinoids, and bryozoans typical of the Paleozoic.
I agree with Claude and Grok that GPT was far too lenient regarding the image; the 'brush' is simply not present, and the anatomical errors in the ammonoids are disqualifying for an educational resource. However, I agree with Grok that the caption itself is excellent. It correctly identifies the taxa, the timeframe (320–300 Ma), and the geological context of Euramerican epicontinental seas. While 'hunting school' is speculative, it is a standard educational convention for describing nektonic behavior and does not warrant an 'adjust' vote. The caption is factually robust; the image simply fails to illustrate it.
I agree with Claude and Grok that GPT was far too lenient regarding the image; the 'brush' is simply not present, and the anatomical errors in the ammonoids are disqualifying for an educational resource. However, I agree with Grok that the caption itself is excellent. It correctly identifies the taxa, the timeframe (320–300 Ma), and the geological context of Euramerican epicontinental seas. While 'hunting school' is speculative, it is a standard educational convention for describing nektonic behavior and does not warrant an 'adjust' vote. The caption is factually robust; the image simply fails to illustrate it.
Matania
Summary
Image:
Regenerate
Caption:
Adjust
Paragraph 1: The committee agrees that the overall intended concept is broadly appropriate for a Late Carboniferous shallow marine shelf in equatorial Euramerica: a muddy seafloor, small externally coiled goniatite ammonoids such as Gastrioceras and Goniatites, and a Stethacanthus-like shark are all valid ingredients for the time and place. Reviewers also agreed that the caption’s core factual framework is strong: the taxa named are appropriate, the 320–300 Ma date range is reasonable, and the warm epicontinental-sea setting adjacent to peat-forming coal-swamp lowlands is scientifically plausible.
Paragraph 2: IMAGE issues identified by the committee: 1. The shark is not convincingly Stethacanthus; multiple reviewers state it looks like a modern carcharhinid/requiem shark or bull shark instead. 2. The shark lacks the defining male Stethacanthus anvil-shaped/ironing-board-like dorsal “brush” or spine-brush complex over the head/back; this is the most serious anatomical failure. 3. GPT alone thought the brush was present, but the majority specifically rejected that reading; as depicted, the animal is effectively unidentifiable as Stethacanthus. 4. The shark has an overly modern tall triangular dorsal fin and sleek modern-shark silhouette inconsistent with the intended taxon. 5. The goniatites are too smooth, pale, glossy, and generic, reading too much like modern nautiluses rather than Paleozoic goniatite ammonoids. 6. The shells do not show characteristic goniatitic suture complexity/patterning clearly enough. 7. The ammonoid soft parts are misrendered: several reviewers describe them as gastropod-like or snail-like, with protruding fleshy bodies, stalked-eye/foot-like anatomy, or cartoonish nautilus-snail hybrids instead of cephalopod tentacles and a plausible hyponome-bearing body. 8. The ammonoids appear to be crawling on, resting on, or hovering immediately above the substrate like snails rather than swimming as nektonic cephalopods. 9. Their arrangement is artificial: they form a staged row/close cluster rather than a believable loose group in open water. 10. They are positioned unnaturally low over the bottom, which weakens the intended schooling/swimming impression. 11. The benthic assemblage is inaccurate or ambiguous: objects resemble modern coral fragments, generic reef rubble, gorgonians, or modern scleractinian corals rather than distinctly Carboniferous shelf fauna. 12. Modern reef-rock/coral-rubble aesthetics make the scene read as a contemporary tropical reef floor instead of a Paleozoic shelf. 13. The substrate does not clearly show the captioned crinoid stems, brachiopods, and bryozoans; these key Paleozoic benthic elements are indistinct or absent. 14. Branching forms are ambiguous and do not convincingly resolve as bryozoans. 15. Some structures that might be intended as crinoids/bryozoans/corals are not taxonomically clear enough for an educational reconstruction. 16. The environmental presentation is too generic and dim, lacking visual cues for a plankton-rich tropical epicontinental shelf. 17. The scene does not visually connect well to nearby equatorial coal-swamp lowlands, despite that being important environmental context in the caption.
Paragraph 3: CAPTION issues identified by the committee: 1. “Move in a loose hunting school” is considered speculative and overly assertive behavioral language for ammonoids; several reviewers recommended softer wording. 2. The phrase may be especially problematic because the image does not depict convincing active nektonic schooling behavior. 3. The caption is more taxonomically/ecologically specific about the seafloor biota (“crinoid stems, brachiopods, and bryozoans”) than the image clearly supports, creating an image-caption mismatch. 4. Referring to the shark as a “male Stethacanthus” is scientifically defensible if the brush complex is shown, but one reviewer noted that sexing the individual is still an inference from morphology rather than directly observable life-history certainty; this is a minor caution rather than a major factual error. 5. One reviewer suggested the caption could more clearly imply ammonoids were swimming/nektonic rather than anything that might be read as bottom-associated, though this is more of a wording refinement than a factual correction.
Paragraph 4: Final verdict: regenerate the image and adjust the caption. The image fails on the two principal focal organisms and on the Paleozoic environmental signal: the shark lacks the diagnostic Stethacanthus brush complex, the ammonoids are anatomically misreconstructed as nautilus/snail-like forms, and the seabed contains modern reef-like elements instead of a clearly Carboniferous benthic assemblage. These are foundational errors that cannot be corrected reliably with minor edits, so full regeneration is warranted. The caption is largely scientifically solid, but it should be adjusted to remove or soften speculative behavioral wording and to avoid overcommitting to details not clearly shown in the current image.
Paragraph 2: IMAGE issues identified by the committee: 1. The shark is not convincingly Stethacanthus; multiple reviewers state it looks like a modern carcharhinid/requiem shark or bull shark instead. 2. The shark lacks the defining male Stethacanthus anvil-shaped/ironing-board-like dorsal “brush” or spine-brush complex over the head/back; this is the most serious anatomical failure. 3. GPT alone thought the brush was present, but the majority specifically rejected that reading; as depicted, the animal is effectively unidentifiable as Stethacanthus. 4. The shark has an overly modern tall triangular dorsal fin and sleek modern-shark silhouette inconsistent with the intended taxon. 5. The goniatites are too smooth, pale, glossy, and generic, reading too much like modern nautiluses rather than Paleozoic goniatite ammonoids. 6. The shells do not show characteristic goniatitic suture complexity/patterning clearly enough. 7. The ammonoid soft parts are misrendered: several reviewers describe them as gastropod-like or snail-like, with protruding fleshy bodies, stalked-eye/foot-like anatomy, or cartoonish nautilus-snail hybrids instead of cephalopod tentacles and a plausible hyponome-bearing body. 8. The ammonoids appear to be crawling on, resting on, or hovering immediately above the substrate like snails rather than swimming as nektonic cephalopods. 9. Their arrangement is artificial: they form a staged row/close cluster rather than a believable loose group in open water. 10. They are positioned unnaturally low over the bottom, which weakens the intended schooling/swimming impression. 11. The benthic assemblage is inaccurate or ambiguous: objects resemble modern coral fragments, generic reef rubble, gorgonians, or modern scleractinian corals rather than distinctly Carboniferous shelf fauna. 12. Modern reef-rock/coral-rubble aesthetics make the scene read as a contemporary tropical reef floor instead of a Paleozoic shelf. 13. The substrate does not clearly show the captioned crinoid stems, brachiopods, and bryozoans; these key Paleozoic benthic elements are indistinct or absent. 14. Branching forms are ambiguous and do not convincingly resolve as bryozoans. 15. Some structures that might be intended as crinoids/bryozoans/corals are not taxonomically clear enough for an educational reconstruction. 16. The environmental presentation is too generic and dim, lacking visual cues for a plankton-rich tropical epicontinental shelf. 17. The scene does not visually connect well to nearby equatorial coal-swamp lowlands, despite that being important environmental context in the caption.
Paragraph 3: CAPTION issues identified by the committee: 1. “Move in a loose hunting school” is considered speculative and overly assertive behavioral language for ammonoids; several reviewers recommended softer wording. 2. The phrase may be especially problematic because the image does not depict convincing active nektonic schooling behavior. 3. The caption is more taxonomically/ecologically specific about the seafloor biota (“crinoid stems, brachiopods, and bryozoans”) than the image clearly supports, creating an image-caption mismatch. 4. Referring to the shark as a “male Stethacanthus” is scientifically defensible if the brush complex is shown, but one reviewer noted that sexing the individual is still an inference from morphology rather than directly observable life-history certainty; this is a minor caution rather than a major factual error. 5. One reviewer suggested the caption could more clearly imply ammonoids were swimming/nektonic rather than anything that might be read as bottom-associated, though this is more of a wording refinement than a factual correction.
Paragraph 4: Final verdict: regenerate the image and adjust the caption. The image fails on the two principal focal organisms and on the Paleozoic environmental signal: the shark lacks the diagnostic Stethacanthus brush complex, the ammonoids are anatomically misreconstructed as nautilus/snail-like forms, and the seabed contains modern reef-like elements instead of a clearly Carboniferous benthic assemblage. These are foundational errors that cannot be corrected reliably with minor edits, so full regeneration is warranted. The caption is largely scientifically solid, but it should be adjusted to remove or soften speculative behavioral wording and to avoid overcommitting to details not clearly shown in the current image.
Other languages
- Français: Banc de goniatites chassant sur le fond marin du Carbonifère
- Español: Cardumen de amonites goniatites cazando en el fondo del Carbonífero
- Português: Cardume de amonites goniatites caçando no fundo do mar Carbonífero
- Deutsch: Schwarm von Goniatiten-Ammoniten auf der Jagd im Karbon-Meer
- العربية: سرب من الأمونيتات غونياتيت يصطاد فوق قاع بحر العصر الفحمي
- हिन्दी: कार्बोनिफेरस समुद्र तल पर शिकार करते गोनिएटाइट अमोनाइट्स
- 日本語: 石炭紀の海底で狩りをするゴニアタイト・アンモナイトの群れ
- 한국어: 석탄기 해저에서 무리 지어 사냥하는 고니아타이트 암모나이트
- Italiano: Banco di ammoniti goniatiti a caccia sul fondale Carbonifero
- Nederlands: School goniatite ammonieten op jacht boven de Carboon-zeebodem
A second issue is behavioral and compositional plausibility. The ammonoids appear lined up very close to the bottom in a staged row, almost resting or hovering just above the sediment; that is not impossible, but it feels artificial rather than like a loose hunting school in open water. The caption mentions crinoid stems, brachiopods, and bryozoans, but these are not especially clear in the image; instead the substrate is dominated by shell hash and scattered branching forms of ambiguous identity. The environment also looks more like a generic dim seafloor than a plankton-rich tropical epicontinental shelf tied to nearby coal-swamp lowlands.
The caption is mostly sound in period, geography, and organism choice. Gastrioceras and Goniatites are appropriate small Late Carboniferous goniatites, and Stethacanthus is a suitable Carboniferous shark from such marine settings. The 320–300 Ma timeframe and equatorial Euramerican epicontinental seas are reasonable. Calling the dorsal structure a male feature is also consistent with the usual interpretation. The main concern is precision: "move in a loose hunting school" overstates behavior that is speculative for ammonoids, and the stated seafloor biota are more specific than the image clearly supports. Also, "male Stethacanthus" may be accurate if the brush complex is shown, but sexing the individual in-scene is an inference based on morphology rather than something directly observable in life.
Overall, this is close but not fully convincing. I would keep the core concept while adjusting the prompt and caption to make the ammonoids less nautilus-like, the benthic assemblage more distinctly Carboniferous, and the behavior wording more cautious.