Modeling
About Archicad's design tools, element connections, modeling concepts, etc.

AI BIM. Can ArchiCAD be ready for that?

Podolsky
Ace
Hi there!

I have decided to start a separate thread about Artificial Intelligence in AEC. I hope here we will able to collect as much information as what happens today in the AI world, show examples of using AI in BIM programs and programs driven by AI, and discuss what architects and other AEC professionals would like to expect from it.

I have a very strong opinion, that these technologies will come to us very, very soon (because AI already exists in many things we are using every day - like search engines or digital photo applications) and this might be a very interesting subject to review.

Please find below a scheme I have prepared, showing how AI for BIM might look like.

Building elements (as we know tools in ArchiCAD) are controlled by placement algorithms, that coming from building classification databases). For example - placing partition walls in the office with the right chosen sound insulation, fire ratings, correct corridors lengths, fire escapes, etc.

I/O engine responsible for Input / Output - but in architectural terms - automatic drawings generation and publishing, remote communication, including communication via e-mails, teamwork, IFC exchange. It is something like a secretary-robot, that supervising the BIM project.

Language analyzer transforming languages into commands, understandable by the system (software). It can understand human language and communicate with him as a chatbot, it’s also can understand different languages - like Python, JavaScript, AutoLisp, etc. Even read IFC (because IFC is a script).

Physical simulations help to improve correct element placement. Simulations shall be 100% on physics (more like physical engines in 3D animation software). Includes loads, earthquakes, heat distribution and loss, fire spread, wind load, radiosity, and photon tracing, similar to Monte Carlo.

Additional block called construction simulations helps to represent the construction process, including delivery, animation of cranes and installation process, construction timeline, and similar.

Each building element has two additional layers - assembly (if it’s a wall, then it might be studs, cover, insulation, and brackets) and behavior. Behaviour is connected to both physical and construction simulations.
68 REPLIES 68
rob2218
Enthusiast
.....but in reality this successful collaboration does not happen - because nobody really wants to collaborate.....

And here in lies the biggest issue. Why you ask (well maybe you don't but I'm going to tell you anyway).
Because NO ONE trust anyone these days.

Before (say...50+ years ago) a contractor TRUSTED that an architect knew W.T.F. they were doing and vice-versa...but today each is out for themselves.

So...with that initial approach to a project, damn the torpedoes and full-steam ahead with YOUR vantage point and don't lose money while yer at it.
...Bobby Hollywood live from...
i>u
Edgewater, FL!
SOFTWARE VERSION:
Archicad 22, Archicad 23
Windows7 -OS, MAC Maverick OS
DGSketcher
Legend
I watched a chilling program about AI the other night. The most reassuring thing in it was the US military can achieve advanced targeting but they have realised that the final decision to pull the trigger should always made by a human. Ok its bit off topic, but I do think we need to walk a careful line between as Braza describes it "Assisted Intelligence" and Artificial Intelligence which is fully autonomous. In the same program they showed how AI was used to in medicine to target Radiotherapy for prostate cancer and was saving Doctors hours of analysis and assessment, but again it was used as a tool to assist the Doctors. I don't mind GS digging into advanced assistance options to take a bigger market share, but we need the option to ignore it when the solution conflicts with our personal objectives e.g. to back reference a comment I made earlier the Designer may prefer the view to a recommendation on solar gain.
Apple iMac Intel i9 / macOS Sonoma / AC27UKI (most recent builds.. if they work)
LaszloNagy wrote:
Can I ask and can someone explain to me what is AI about anything in this video?
This is simply an algorithm with many parameters and variables. If this is AI, then most GDL objects are AI, as they generate countless variations of geometry based on user input and parameters.
My understanding is that AI makes decisions on its own which looks like human intelligence.
I don't see anything in this video that does that, anything that is not determined by a human programmer.

I think this would be AI if the program would place the variant it thinks is the best without asking a human and it would be confident in its decision. So, the AI would have to be able to have its own viewpoint and would have to be able to form its own opinion.
Also, my understanding is that AI involves Machine Learning. I don't see any learning on the part of the program in this video.


Emre wrote:
Just to add to what Podolsky is saying, here's a quick video of the tool in action:
.......
Whether or not this is well implemented in Revit I don't know, but from a first look, it seems like an interesting tool that would be certainly of use to bigger practices, and perhaps even small ones. Not saying this is the most important tool needed in AC, but it is a very good example of AI implementation.



It is not AI (...in the strictest sense of the term).

It is at best,...."AI-like". (while, not really)

There's a lot of conflation in this discussion between the expressions like "AI" (Artificial Intelligence, which is a whole separate, deeply involved field all on its own in Computer science technology, and which does involved aspects like Machine learning, Evolutionary algorithms, Genetic algorithms, etc)), "AI-like" algorithms (which are NOT AI in the strictest sense, but perform certain functions that may seem to mimic parts of of how AI operates) and finally "AI-assisted algorithms" (also not AI, but closer to AI than "AI-like" algorithms, but also in themselves mimicking certain functions of AI-functionality in limited fashion as part of larger algorithms or programs that have nothing at all to do with AI).
And admittedly and arguably there's probably a lot of cross-over between what would be considered "AI-assisted" and "AI-like"

That's why I think the whole discussion about AI in that other thread is woefully misguided and wrong-headed because most of what they're talking about there have to do with the latter two (AI-like algorithms and AI-assisted algorithms), and not strictly "AI" itself.

Nobody is going to create a strictly speaking AI algorithm or AI program to replace the functions of what architects do at a holistic level any time soon because what architects do can't be boiled down to just data and numbers as can be done in other fields.
You're not going to have automated drafters working at architecture shops the way you have automated tellers at banks replacing human tellers, or automated checkout stations at supermarkets replacing human clerks because most of what constitutes the functions of those kind of positions easily replaced by algorithm-driven auto-teller or auto-cashier machines are routine, rote-based, repetitive, data-driven duties and functions that involve little or not problem-solving, creativity, and independent decision-making on the level that you have with architecture.
(and to be really strict here, the algorithms and programs that drive those automated cash machines or auto-checkouts at the supermarket are NOT "AI" themselves at all even in the most liberal sense, and at best are only partially "AI-assisted". If that even)

Even taking it farther to more complex sort of algorithms and programs that perform really high-skill tasks and jobs. Autopilots in both the automobile and airplane sense (which we all can agree are relatively very complex tasks and high-(mental)concentration sort of things for humans to do), the algorithms in those are still very much data driven and predicated on operating within a very well-defined set of parameters and rules that don't make or leave much room for the sort of innovation and creative thinking (even in the simulated sense) that defines how we understand "AI" to work.
There's a reason, beyond just liability and legal considerations why the likes of Tesla still require a fully qualified human driver sitting behind the wheel of their Autopilot-driven cars and why even despite the fact that planes nowadays can pretty much take-off, fly intercontinental and land themselves all on autopilot, they STILL require at least TWO human pilots in the cockpit for the entire flight.

Ironically, one of the best examples of what a basic "AI" algorithm would be is the kinds that have been with us almost since the inception of computer programming.

Chess programs.
Present them with unexpected scenarios with each iteration and move, in trying to prevent them from accomplishing their goal, and the program devises an optimal response (based on analysis, calculation, weighing of probabilities and possible scenarios, on and up to a level most humans can't. i.e. still data-driven) from an set of multiple options available to it.
Wash. Rinse. Repeat until enemy King is dead.
That, on a basic level is how we as humans think.

Adapting to each changing set of circumstances especially from unforeseen or unforeseeable outcomes.

Most "AI-assisted" or "AI-like" algorithms are not capable of that, nor designed to do that, and it's what strictly speaking AI technology study tends to focus on, including machine learning.

There's a whole lot of complexity and mystery as to how the human brain works and how we think, adapt to situations, solve problems, create and design - a lot of which we still don't even understand fully.
The field of Artificial Intelligence barely seeks to create a basis for simulating this on a very rudimentary level, before you even begin to account for more esoteric aspects of cognition like morality, philosophy, "faith" or spirituality and the like.
jl_lt
Ace
another case: You can have all the variables and inputs from the client for any given project, and a computer could churn out millions of variations according to those variables, all of them viable. But what if i, as an architect, decide that the original premises, inputs or "what the clients think they want" is wrong? Great projects have been created because the architect had a different (and ocassionally better) vision as what was originally stated. Let me know when a computer can do that.
Laszlo Nagy
Community Admin
Community Admin

Lots of valuable insights here in these last several posts, guys.

I feel like AI is like BIM was a few years ago and still is for some - a catchphrase or a word to throw around to get more business.
There were/are people who said/say they are doing BIM, where in reality, maybe they were/are only doing some 3D modeling and calling that BIM.
Now people are doing Generative Design, using complex and smart algorithms, and so on, and they are calling that Artificial Intelligence.

These hype terms are being "abused" a lot.

Loving Archicad since 1995 - Find Archicad Tips at x.com/laszlonagy
AMD Ryzen9 5900X CPU, 64 GB RAM 3600 MHz, Nvidia GTX 1060 6GB, 500 GB NVMe SSD
2x28" (2560x1440), Windows 10 PRO ENG, Ac20-Ac27
Podolsky
Ace
Today a lot abuse and hype happens. That noise, produced by hypers disturbing to concentrate on the subject, that in this case is purely mathematical. Very possible that common mental level of humanity will continue fall down when AI technologies will continue rise. This is another, "idiocracy" - style scenario.

Another example of AI:
https://www.jpost.com/arab-israeli-conflict/gaza-news/guardian-of-the-walls-the-first-ai-war-669371
Anonymous
Not applicable
Bricklyne wrote:
Ironically, one of the best examples of what a basic "AI" algorithm would be is the kinds that have been with us almost since the inception of computer programming.
Chess programs.
Agreed. And also talking about "Real AI", I recall reading a scientific paper sometime ago that described an AI experiment where a certain number of individual AI systems (based on neuronal networks) were only "allowed" to communicate between themselves. And surprisingly the scientists noticed that they had developed their own language!
Anonymous
Not applicable
Just tried to find the original documents... so it was back in 2017.
Actually the paper was published by 2 Facebook researchers.
I found a recent article revisiting the subject, and looks like, the "incident" has been used and abused a lot since then. The recent article says that this incident is actually very common in these kind of experiments involving AI and end-to-end negotiations. And for now doesn't bring any concern for researchers. But it shows that machine learning can extrapolate human language to achieve "personal" goals.

The bots were given training data sets and let these bots negotiate with each other. While the original data sets were in English, the bots ended up taking certain phrases from them and put them out based on numerical representation and by analyzing the outcome. The engineers decided to stop the simulation after it started producing results that were degenerated and looked strange.
Podolsky
Ace
LaszloNagy wrote:
Now people are doing Generative Desing, using complex and smart algorithms, and so on, and they are calling that Artificial Intelligence.

These hype terms are being "abused" a lot.
Funny enough, but Graphisoft also was playing similar game when introduced new stair tool using AI. Intension is very good - but currently stair still have many small bits to be fixed / improved, because currently opinion of this tool is doubled - some people like it, some hate it. About AI in stair tool nobody even think today.
Podolsky
Ace
What I also would like to say - about using AI by army (and especially by IDF in the last war). Of course, as we know, all computer technologies came from US army. Telecommunication was developed by US army during Second World War, later it turned into AT&T, that developed UNIX and C language. Because of that we all now have personal computers. Programming language - as mathematical concept - was created by army. Early versions of internet (later it moved into universities, where was created www concept). Now it looks like the most advanced AI technologies are developed by US and Israel army. And we will get them later too.
So why we need to destroy, explode, kill first - and after build? Just imagine - the whole network of underground tunnels in Gaza was controlled via AI. But as I know - they still don't have proper water supply and collecting rainwater on the roofs.
If such a technology to use to control and develop underground communications in the cities? Water, electricity, internet, underground trains...