A dream of pink pastel colors from Reform. Peach and pale pink are the exact shades of the classic BASIS collection. With the mission to offer kitchen design at affordable prices, Jeppe Christensen and Michael Andersen founded Reform in 2014. With 13 showrooms around the world, from Germany to Denmark to the USA, you get the opportunity to see the kitchen designs.
Check out this post on Instagram
Low-angle photograph of a yellow concrete skyscraper in Moscow, taken by Maria Krasnova.
AI-supported building research on historic buildings is the new discipline that is kicking the dust off archives and digitizing mortar cracks. Between algorithmic precision and architectural empathy, the industry is exploring the great promise: What can artificial intelligence really do when it follows the traces of the past? And are Germany, Austria and Switzerland ready for a revolution in dealing with built heritage?
Traditional building research used to be a craft of magnifying glasses, folding rulers and dusty archive boxes. Today, however, the tools have changed radically. Artificial intelligence not only sifts through sheer endless amounts of construction plans, damage reports and building documentation, but also recognizes patterns that remain hidden to the human eye. Although this development has not yet been implemented across the board in Germany, Austria and Switzerland, pilot projects are on the increase. Whether digital damage mapping of Gothic vaults or semantic analysis of building phases – AI is becoming the standard tool for the architectural research of tomorrow.
The wave of innovation is rolling with mathematical precision and disruptive force. Traditional photogrammetry is no longer enough. AI-supported algorithms analyze gigabytes of image material, extracting information on crack widths, material decay or structural changes. Laser scanning and 3D photogrammetry are combined with deep learning workflows that not only generate models from point clouds, but also intelligent diagnoses. Historic facades thus become a data set that remains legible for centuries – and at a level of detail that amazes even veteran conservationists.
However, it is not just the technology that is changing, but also the approach. Building research is becoming a digital detective game in which algorithms act as trackers. They recognize hidden windows, reconstruct conversions and identify the handwriting of individual Baumeisters. In Zurich, for example, an AI system analyzes the building history of old town houses on the basis of historical plans and current sensor technology. In Vienna, AI-based damage analyses are integrated into the planning of renovation measures in order to use resources in a targeted and sustainable manner.
All of this is putting the discipline’s self-image to the test. Will building research in future be more a question of data models than experience? Who controls the interpretation of algorithms? And how can we prevent AI-based diagnoses from becoming dogma for restoration decisions? The industry is grappling with these questions – not least because technology does not make people superfluous, but forces them to acquire new skills.
The digital search for evidence opens up unimagined possibilities, but also new gray areas. What is considered an objective analysis is often the result of algorithmic assumptions. Who decides which training data the AI receives? Where is the distinction made between patina and structural damage? Building research is thus becoming an arena for the technical, ethical and cultural struggle for the authority to interpret the built heritage.
The innovative power of AI-supported building research is not a sure-fire success, but a product of intensive research, courageous pilot projects and permanent technical evolution. While traditional methods are based on careful observation and manual documentation, the new tools rely on automated analysis, data synthesis and intelligent pattern recognition. The triad of 3D laser scanning, machine learning and semantic modeling forms the backbone of this technical revolution.
In Zurich, for example, the facades of historic buildings are scanned in millimetre resolution and examined by AI systems for damage, material changes and construction details. In Germany, databases are being created that use millions of building drawings and damage reports to train AI applications. The algorithms learn to distinguish between crack patterns and age-related material deterioration and predict the long-term development of structural damage. The result: early detection, more precise diagnoses and targeted renovation strategies that conserve resources and preserve historical substance.
However, the technical complexity is challenging the industry. The integration of heterogeneous data sources – from historical plans to current measurement data and sensor information – requires interoperable platforms and robust data standards. If you want to be at the forefront here, you not only need structural engineering expertise, but also knowledge of data science, machine learning and digital modeling. The interfaces between architecture, IT and heritage conservation will become a key skill for the next generation of building researchers.
The role of open source and open data is growing. In Austria, for example, parts of the building research data sets are being made publicly accessible in order to broaden the training base for AI applications. At the same time, new formats such as semantically enriched BIM models are being created that link historical information with current building conditions. The digital twin of a listed building thus becomes not just an image, but an adaptive knowledge repository that can accompany restoration and use for generations to come.
But not everything shines. The quality of the AI results depends on the quality of the data and the transparency of the algorithms. Black box models, a lack of documentation and dependence on proprietary software pose new risks. Construction research must learn not only to use technology, but also to question it critically. Only those who understand how AI works can interpret its results in a meaningful way – and prevent algorithmic artifacts from becoming the new truth.
Sustainability in building research has long been a marginal topic, somewhere between energy-efficient refurbishment and careful material selection. AI-supported methods are now opening up a new playing field. The data-based analysis of historic buildings makes it possible to plan renovation measures with pinpoint accuracy – with the lowest possible use of resources and maximum protection of the substance. In Germany, for example, AI-based damage predictions are used to renovate those components that are actually at risk, while other areas remain untouched. This not only saves costs, but also reduces the ecological footprint of heritage conservation.
Adaptive use, circular economy and energy optimization are made tangible by AI. In Switzerland, algorithms analyze the potential of old buildings for contemporary uses – office, residential, cultural – and simulate the effects of various conversion options on energy requirements, daylight and room comfort. The combination of historical data, real-time measurements and digital models enables sustainable usage strategies that keep the built heritage alive instead of preserving it as a museum.
However, the sustainability of the digital tools themselves is also up for debate. The energy requirements of large AI models, the long-term archiving of digital data and the issue of digital obsolescence are unresolved challenges. Anyone relying on AI-supported construction research today must also think about the lifespan and accessibility of the digital results. New standards, open formats and sustainable infrastructures are needed here – otherwise the digital treasures of the present risk becoming digital ruins tomorrow.
Another field: the targeted reuse of historical materials is supported by AI-supported material analyses and traceability systems. In Austria, for example, bricks, wood and metals are classified using machine learning and recorded in material databases so that they can be specifically reused in later construction projects. The circular economy is thus also reaching monument preservation – and making historic buildings a role model for sustainable architecture.
The challenges are considerable, the opportunities enormous. AI cannot guarantee sustainability, but it does provide the tools to make informed decisions. However, the responsibility to use these tools wisely remains with humans. Those who see it as a mere efficiency machine will forfeit the cultural value of the built heritage. Those who see them as partners in a long-term, sustainable dialog can lead historic buildings into a resource-conserving future.
The demands on building researchers, architects and conservationists are changing fundamentally. In addition to the classic canon of building history, materials science and restoration techniques, digital skills are taking center stage. Anyone who wants to play a part in building research in the future will have to speak the language of algorithms, structure databases and understand the logic of machine learning – without losing their architectural judgment.
Academic programs in Germany, Austria and Switzerland are slowly responding to this development. The first degree courses are integrating modules on data science, AI and digital modeling into architecture training. But skepticism still dominates. Many practitioners fear the loss of craftsmanship, the feeling for materials and space that comes from decades of experience. They see AI as a cold tool that replaces the genius loci with statistics. The debate is emotional, often characterized by misunderstandings – but necessary to prepare the profession for the future.
Technical know-how alone is not enough. The ability to critically evaluate AI models, check data sources and recognize algorithmic bias is becoming a key skill. Those who do not question the results of AI run the risk of perpetuating errors and promoting cultural misinterpretations. The new generation of construction researchers must therefore not only be able to think digitally, but also critically and interdisciplinarily.
Collaboration is also changing. AI-supported building research is teamwork: computer scientists, architects, building historians and conservationists work together on the digital model. Traditional disciplines are merging and hierarchies are blurring. If you don’t want to lose touch, you have to build bridges – between software development and building research, between laboratory and construction site, between digital simulation and real intervention.
The job profile is expanding. New roles are emerging: Data Curator, Digital Heritage Specialist, Algorithmic Consultant. Building research is becoming more international, more networked and faster. Anyone who wants to shape it needs the courage to experiment, the desire to try new things – and the willingness to constantly question and expand their own knowledge.
Like every technological revolution, AI-supported construction research is a field full of ambivalence. The euphoria about new possibilities meets skepticism towards algorithms that are not yet known for their cultural sensitivity. Critics warn against the over-technicalization of heritage conservation, the reduction of historic buildings to data points and the danger of cultural diversity being leveled by global software standards. The discussion is necessary – and has long been part of the international architectural discourse.
Global initiatives such as Open Heritage Data, Digital Preservation and Heritage BIM are influencing regional practice. In Switzerland, for example, open platforms are being created on which restoration data, damage images and building phase models are made accessible for research and practice. International networking ensures the transfer of knowledge, but also the risk of local characteristics being lost. Anyone conducting research with AI must be aware of the tension between standardization and cultural diversity.
Visionary voices see AI-based building research as an opportunity to save the built heritage for the future in the first place. They argue that only through intelligent digitization can the wealth of information be preserved, analysed and made usable for future generations. Others warn against a digital transformation that would result in the loss of authenticity. As is so often the case, the truth lies somewhere in between.
The question of control and transparency is particularly controversial. Who decides which data flows into the AI? Who owns the digital models? How can we prevent commercial interests or technocratic bias from determining the interpretation of built heritage? Building research must develop new governance models, demand open standards and retain control over its digital tools.
One thing is certain: AI will not replace building research, but it will radically change it. The industry has a choice: either it actively shapes the change – or it will be overtaken by the algorithms of others. What remains is the responsibility to combine technical potential with cultural sensitivity and professional expertise. The future of construction research is digital – but it remains a question of attitude.
The digital search for evidence using AI is more than just a technical trend. It is a paradigm shift that is renegotiating the relationship between people, buildings and knowledge. In Germany, Austria and Switzerland, the signs are pointing to a new beginning – but also to debate, experimentation and critical reflection. Those who take advantage of the opportunities offered by AI without losing sight of the peculiarities of the built heritage can forge new paths towards sustainable, sensitive and in-depth building research. The future belongs to those who have the courage to combine traditional tools with digital methods – and not leave cultural memory to the algorithm.
attacked by wolves" (1844
Do you have to believe everything you don’t know? The saying about Corot’s oeuvre is well known: “Around 3,000 works by Corot’s hand have survived, 6,000 of which are in America.” But how can you recognize an original, and what is actually “original”? Is original only what is “old” or what comes from the “master”? How […]
Do you have to believe everything you don’t know? The saying about Corot’s oeuvre is well known: “Around 3,000 works by Corot’s hand have survived, 6,000 of which are in America.” But how can you recognize an original, and what is actually “original”?
Is only what is “old” or what comes from the “master” original? What about different versions of a successful and therefore often re-created theme by the same painter (fig. 2), what about joint works by different painters, how are workshop copies to be assessed? What is the situation when (old) masters copy other (old) masters? Testimonies of art history, which often concern us more than the verified, “genuine” works, are to receive a brief, illuminating treatment here. It is comforting to know: “Si duo faciunt idem, non est idem!” (“If two do the same thing, it is not the same thing!”), because many doubts can be verified today. This statement also describes the cause of many disputes that have arisen since works of fine art have been so highly prized. Fortunately, especially in the case of paintings, there are many ways of recognizing them. Natural ageing can be checked, wrinkles, craquelure, cracks in the stretcher and stretcher frame, soiling, yellowing, colorless plastering cannot easily be imitated. The amount of restoration work carried out is also often an indication of authenticity, as supposedly valuable paintings have often been treated, i.e. have a real history. “Habent sua fata pinaces” – paintings have their fates. The measures carried out at different times also reflect a history of the restoration methods used at different times.
The signature is the fetish of all art historians and art dealers. Once it is on the painting, it is blindly trusted. This trust is not always appropriate. The signature on a painting does indeed authorize it through the artist’s signature. Since the whole world believes in the authority of the signature, it is often manipulated, usually with dishonest intentions. Healthy doubt can save money in the case of an intended purchase or loss of reputation in the case of an intended attribution. In addition to the originals signed by the artist, later attributions also led to signatures being subsequently painted on. A bizarre way of dealing with one’s own signature is provided by Salvador Dalí, who signed blank canvases and then did not paint them, thus turning the pictures later painted on these canvases by other painters into signed Dalís. He also signed Dalí works by his friend Antoni Pixtot, who had painted them in Dalí’s style in Dalí’s villa when Dalí was no longer able to execute them himself due to physical limitations. Picasso signed one of his paintings with Boldini because the sitter said she would have preferred to have been painted by him. And many a painting that was painted by Picasso and Braque at the same time in the studio and which resembled each other so closely was signed by one or the other as a joke, also because the painters no longer knew which one was painted by whom. But even works definitely painted by the artist, which years after their creation are no longer identified with them and are banned from their catalogs raisonnés, cannot be legally recognized – especially since they were authorized by the artist. The animal painter Norbertine von Bresslern-Roth (1891-1978) was also very critical of her results: She did not want to release all her paintings into the public domain. She cut up paintings that did not meet her standards and disposed of them in her own dustbin. Once word of this practice had spread among her neighbors, who took the paintings she had destroyed out of the garbage and restored them, she preferred to use other people’s garbage cans in the vicinity of her home, where she only put parts of destroyed paintings.
Re-dedications sometimes took place in studios shared by two or more painters, such as the Alsatian Eck brothers, where one brother finished painting the deceased brother’s picture, painted over his signature and then signed it himself. Overpaintings of the names of less highly traded painters with the names of more highly paid painters are known. For example, a painting by Jaques Fouquière was rededicated as a Joos de Momper. However, original signatures were also removed: this was often done unintentionally in the case of works by David Teniers the Younger, where the value-creating signature is usually on the final varnish, which was very often lost when the varnish was removed.
The loss of the signature during restoration is often the reason for “reconstructing” such a signature, which is prohibited under copyright law. In most cases, these repainted signatures look less reliable than the original ones and are the reason why an original painting is disallowed, even though “only” the signature is fake. If such newer signatures lie on a surface that has been craquelured through because it is older, it is possible to recognize with a microscope that it is a later addition.
Forged signatures, on the other hand, are not always deliberate misrepresentations; the copyists often took the view that the signature was a creative device of the painter and that a copy was only complete through the carefully copied signature. The painter Dirk Huisken in Celle, who was close to the “Brücke”, made copies of various painters for his private studio in order to study their technique and deliberately added altered signatures to them. He turned the name Picasso into “Picolino” so as not to gain the reputation of a forger after his death.
It should not be forgotten that in the past it was not considered dishonorable to train oneself on old masters and copy them, on the contrary. It had nothing to do with “forgery” in the modern sense. People imitated the master to show that they were on a par with him. Rubens, for example, also copied diligently, even if not always exactly. Jacob Burckhardt writes about Rubens as a copyist: “When Rubens copied not for the Duke of Mantua, but for himself, in color or drawing, it was done in the most liberal manner, and sometimes as if he had wished to show the past masters how they should actually have begun.” The Rubens copy of Titian’s “Adam and Eve” (Prado, Madrid, fig. 1) also confirms this statement. Either way, these copies are certainly “genuine Rubens”.
Read more about originals, forgeries and copies in RESTAURO 3/2017.
