The photo shown to artificial intelligence revealed the weak points of the collapsed building: It said disaster is coming

BEUN Geomatics Engineering Department Lecturer Prof. Dr. In October last year, Şenol Hakan Kutoğlu, together with his software team, developed artificial intelligence-supported software within the university, which measures the earthquake resistance of buildings by working in the 'rapid building assessment' format developed by the Federal Emergency Management Agency (FEMA), known as the American AFAD.

Using the software, the earthquake resistance of a building can be determined in accordance with the 'quick building evaluation' format, with the support of artificial intelligence, through the photo and location information of a building, in order to inform the user. The software can be accessed at 'yapianaliz.beun.edu.tr'.

'ARTIFICIAL INTELLIGENCE FOUND MANY GEOMETRIC DEFICIENCIES IN THE BUILDING'

Prof. stated that he had the software we developed examine the building that collapsed in the Maltepe district of Istanbul, using the location information and photographs of the building taken 9 months ago. Dr. Kutoğlu said, "You enter the address or location of the building into this software, and the program automatically calculates the values and seismicity level of the ground where the building is located. In addition, when you upload a clear photo of the facade of the building, it calculates the geometric weaknesses of the building using artificial intelligence. Artificial intelligence can automatically scan whether there are short columns in the building, whether there is a protrusion in the building, whether there is tension, whether there is a soft floor. We decided to evaluate this building after it collapsed, and we found a photo of it from January 2026. When we gave this to the software, the software "The ground is medium-solid, which is not a bad ground. Since it is in Istanbul, the seismicity level was medium-high. However, in the visual evaluation of the building, it found many geometric deficiencies in both vertical and horizontal areas and compiled it into a report. When we looked at the report, the building was classified as a very high risk building with a rate of 95 percent."

'THE BUILDING GOT 0.5 POINT OUT OF 3'

Explaining the deficiencies detected by the software in the building, Kutoğlu said, "It detected that the land settlement of the building was prone, that there were wide openings in the building, detected soft floors, and found short columns. The artificial intelligence also detected wide openings and parapets. It detected a heavy mass structure on the upper floor. Since all these brought negative points to the building, the building could receive a score of 0.5 out of 3. This is not a final evaluation, of course, by taking this evaluation into consideration, our citizens can decide whether to go for a more comprehensive engineering examination. This building is already artificial in appearance." "The intelligence also detected that the building has a very weak appearance. As you can see, from the photo of the building 9 months ago, the artificial intelligence detected both the ground condition and the shape of our building. It was determined with the help of artificial intelligence that it was already a weak building. It may be an exaggeration to say that the building was weak, but it was determined as a result of the visual scan that the building was a very weak building."