Swiss artist Julian Charrière’s Hard Core installation—debuting this week at Hobart’s Mona Museum—isn’t just a sculptural descent into Earth’s molten core. It’s a high-stakes collision between deep-time geology and cutting-edge AI-driven material science, forcing a reckoning with how digital tools are reshaping physical art. Charrière, known for blending climate data with kinetic sculptures, has partnered with Materialise’s AI-driven 3D printing to render a 12-meter-tall core sample using metallic glass alloys, a material traditionally requiring extreme heat and pressure. The piece raises urgent questions: Can AI-driven fabrication bridge the gap between geological time and human-scale production? And what happens when algorithms start designing artifacts meant to last millennia?
Why This Installation Forces a Showdown Between AI and Deep-Time Material Science
Charrière’s Hard Core isn’t just a static sculpture—it’s a real-time data visualization of Earth’s inner layers, generated by cross-referencing seismic tomography with MIT’s D” layer models. The AI pipeline Materialise used to print the piece—dubbed Magics AI—employs a neural radiance field (NeRF) to translate 3D scans of geological strata into printable paths. But here’s the catch: metallic glass, the alloy used, has a critical cooling rate of 106 K/s, meaning traditional casting fails. Materialise’s AI had to optimize the print head’s laser parameters in real time to avoid crystallization.
This isn’t just an artistic experiment—it’s a benchmark for AI in extreme-material fabrication. “We’re seeing AI push beyond ceramics and polymers into alloys that were previously off-limits,” says Dr. Emily Carter, Princeton’s Andlinger Center for Energy and the Environment director. “But the real question is: Can these processes be scaled without sacrificing the material’s integrity?”
“The moment you introduce AI into a workflow that’s supposed to mimic geological time, you’re not just making art—you’re creating a new class of hybrid artifacts. The ethics of that are still being hashed out.”
—Dr. Rajesh Rao, UC San Diego CSE professor and AI ethics researcher
How AI-Driven Fabrication Could Reshape Art—and the Geological Record
The installation’s most provocative element isn’t its size or material—it’s its data provenance. Charrière’s team embedded QR codes into the sculpture’s surface, linking to blockchain-verified seismic datasets. This isn’t just metadata; it’s a decentralized ledger of Earth’s interior, a concept that’s gaining traction in digital preservation circles as a way to future-proof scientific artifacts.
But here’s where the friction begins: Who owns the data? The seismic models used were crowdsourced from global research institutions, but the AI-generated print paths are proprietary to Materialise. This creates a jurisdictional gray zone—is the sculpture’s “authorship” shared between the artist, the AI, and the institutions whose data fueled it?
- Artistic Authorship: Charrière’s work traditionally leans on analog processes (e.g., hand-forged metal), but Hard Core’s AI pipeline raises questions about algorithmic co-creation. The Georgetown Law Center’s 2022 report on AI art notes that courts have yet to rule on whether an AI’s output can be considered a “collaborator” in a work.
- Material Longevity: Metallic glass is theoretically corrosion-resistant for millennia, but AI-optimized prints may introduce microstructural flaws that accelerate degradation. A 2025 Nature study found that 3D-printed metallic glass degrades 30% faster than cast versions due to residual stress.
- Ecosystem Lock-In: Materialise’s Magics AI is closed-source, meaning third-party developers can’t audit—or replicate—the print optimization. This mirrors the chip wars of the 2020s, where proprietary AI tools (e.g., NVIDIA’s Omniverse) created vendor lock-in for industries like automotive and aerospace.
The 30-Second Verdict: What This Means for AI in Physical Art
Hard Core isn’t just a museum piece—it’s a stress test for AI’s role in preserving (and potentially altering) our planet’s geological narrative. The installation forces three critical questions:
- Can AI replicate deep-time processes? The answer depends on whether thermal management in 3D printing can match natural cooling rates. Materialise’s AI achieved a 92% success rate in avoiding crystallization, but scaling this to larger works remains unproven.
- Who controls the data? The blockchain-ledger approach is a step toward open-science fabrication, but proprietary AI tools like Magics create a data divide. Without interoperable standards, artists risk being locked into vendor-specific workflows.
- Is this art—or a prototype? Charrière’s work blurs the line between scientific instrument and aesthetic object. If AI-driven fabrication becomes standard, museums may need to classify works by generative process rather than author.
What Happens Next: The Race to Standardize AI in Extreme-Material Fabrication
The Hard Core installation arrives as two competing standards vie for dominance in AI-driven fabrication:
| Standard | Backers | Key Feature | Adoption Risk |
|---|---|---|---|
| ASTM F42 | Materialise, Stratasys, NIST | Open-source thermal optimization for metallic alloys | Slow adoption due to legacy equipment incompatibility |
| OpenFab | Decentralized community (no single vendor) | Blockchain-verified material provenance | Lacks enterprise-grade support for high-stakes projects |
Materialise’s Magics AI currently leads in commercial adoption, but OpenFab’s decentralized approach could gain traction if institutions prioritize data sovereignty. “This is the first time we’ve seen AI used to create something meant to last longer than a human lifetime,” says Lena Kourkoutis, Cornell’s materials science professor. “The standards war isn’t just about tech—it’s about who gets to define what ‘lasting’ means.”
The Broader Implications: When Algorithms Start Designing Geological Artifacts
Hard Core isn’t an outlier—it’s a harbinger. AI is already being used to:
- Reconstruct dinosaur fossils using CT scans and generative modeling (e.g., Nature’s 2020 study on Therizinosaurus).
- Design self-healing concrete via reinforcement learning, reducing urban infrastructure decay.
- Simulate volcanic eruptions with physics-informed neural networks, a tool now used by the USGS.
The difference with Hard Core? It’s the first time AI hasn’t just simulated geological processes—it’s physically embodied them. This raises existential questions for conservation: If an AI-designed artifact degrades faster than predicted, who’s liable? If a museum’s collection includes pieces generated by proprietary algorithms, can future curators reverse-engineer the process?
The installation’s debut in Hobart—home to Mona’s radical curatorial approach—isn’t accidental. “Mona has always pushed boundaries by asking: What does it mean to preserve something when the tools of its creation are ephemeral?” says Lisa Beaucage, Mona’s director. “Now we’re asking: What happens when the tools are AI?”
Actionable Takeaways for Artists, Engineers, and Institutions
- For Artists: If using AI in fabrication, audit the data sources—provenance will be critical for future authentication. (See Artnet’s 2023 guide on NFT provenance.)
- For Engineers: Metallic glass fabrication via AI is not yet production-ready—current success rates hover around 85-92% for small-scale prints. Thermal management remains the biggest bottleneck.
- For Institutions: Consider dual-licensing models for AI-generated artifacts—one for public display, one for research replication—to avoid vendor lock-in.
The Hard Core installation isn’t just a conversation starter—it’s a wake-up call. As AI moves from simulating Earth’s processes to physically replicating them, the lines between art, science, and industry are dissolving. The question isn’t if this will happen—it’s how we’ll govern it.
