AI- umělá inteligence - interview for Technopolis
AICCS - Artificial Intelligence for the Cultural and Creative Industries
Definition of AI:
For the purpose of this study, Artificial Intelligence is defined as encompassing a set of purpose-driven specific systems able to process information from large datasets and to make predictions and/or take actions and improved decisions at great speed; simulating human-like cognitive functions such as learning, understanding, reasoning, interacting, problem-solving or decision-making. This definition refers to narrow AI systems, as opposed to a general AI system, which would be truly autonomous and able to self-improve independently from humans – a milestone that has not yet been reached.
Application of AI
● Can you mention examples of AI technologies being used in the CCS and in your specific sector (in this case architecture)?
I think we should avoid using term AI until we don´t have standards for levels of AI maturity. It could cause confusion in future, particularly false expectations etc. I prefer to talk more about self-learning or advanced algorithms, but for the purpose of this text I stay with AI. In terms of architecture, we can talk about AI as a tool and AI as a technology that could have an impact on architecture. Though both things are at the very beginning of development and I am waiting for any substantial software improvement. There are also specific applications on construction sites. I heard about smart cameras they can check safety provisions on construction site ,and of course not only this, but that is far distant from architect´s core of work.
● How is AI helping to create content?
Create content should be crucial feature of AI. But AI should not replace human creativity. Still, in many cases, such a future is inevitable. Already today we can listen AI composed music. The same stylized sterility we can expect from the building design. And buildings are not possible to simply switch off. I am not afraid about creative architects. In good hands AI could be a great tool. I am afraid about market inquiry that such “so called creative '' algorithms could be very attractive and profitable for the software developers.
● How is AI helping to promote content?
Once content is created,promotion by using AI will not have limits. But there is a hidden tricky problem which has been continuously revealed by the time of using BIM. With raising possibilities on the side of deliverables are decreasing client´s capabilities on decision-making. Clients are claiming never-ending changes and more enhanced presentation because in their opinion, it is feasible. And then changes continue to the construction site. It concerns claims for variants also. It could be said it is the right place for AI but I don’t think so, that randomized design is what we understand as a design option.
● Reaching wider audiences, in particular younger digital native audiences
AI and all new technologies are a natural matter of interest to young audiences. Particularly in design. Outcomes, deliverables are pleasing. But designing is not making pictures only. Designing is a tool to produce products. It should be effective, the shortest way to explain rightness of concept and then realize construction. It could be task of AI to keep and regulate design in proper and desired limits. It could be also the tool for client to distinguish the functional concept from the pleasing one.
● Enhancing the user experience
Today, from distance, from advertisement, it looks that working with architecture software is so smart and intuitive, but in fact, it is quite the opposite. It is still hard work where mistakes are not forgiven. There are a lot of issues where AI could take a right place like checking norms and automating project documentation in terms of legal rules for a building permit. It is crucial to keep in mind that AI must facilitate designer´s work and not replace him with a position of somebody who is working on “AI ideas”. Seeing the maturity level of current software, we are dealing with errors and imperfections that could have been solved back in ten years. From that perspective, the AI is far behind the horizon. Today, software updates are coming with new functions, new tools providing wider scope of work but updates are rarely improving existing scope of work. In practice it means, that architects are doing more and more work and client´s claims are raising. I wish AI processes would be helpful, saving time, saving energy and optimising what we are already doing.
● How is AI helping to optimise business processes?
Going into detail from previous question. AI could automate deliverables of project documentation. Check norms and regulations. Facilitate sustainable or circular approach from preliminary design. Interconnection with AI systems of state agencies and building authorities would speed up processes of permitting.
● How is AI helping to monetise content and ensure sustainability of business models?
It would be possible to say when we know the cost of AI processes.
● Which AI technology/solution do you expect to spread more in the next five years?
Architectural sector is specific because to many diverse professions are involved and every design is completely tailored to the concrete place, or – should be. I do not expect that Ai technology in our sector will have the same quick development as in medical, financial, IT or mechanical engineering sectors. In medical sector there is a huge claim on public health and mistake avoidance, other sectors having better determined methods.
If I would go back to Ai for architect´s use, I wish I see a spread of more open-sourced solutions and better offers on the market. Such development is already possible to see on visualisations and generic 3d modelling. It is due to wider audience like an advertisement design, film making. AI solutions could also facilitate BIM models and remove dependency on one platform that is currently, in a certain sense, abused by software developers. Probably this is also the reason why it will not occur soon.
● What kind of challenges do you foresee in the application of the use cases you mentioned in your sector? (technical, skills, financial, collaboration)
The following challenges could be summa of the above mentioned. It is about the direction of the next development. If AI will facilitate work or cause additional work. Risk based on overwhelming outputs making more complicated decision-making on the client's side. It is continual rising dependency on a software platform if one AIwill not understand to other AI purposely. It could be a general apprehension from the position of software developers they need to keep architects in expectation for a new update to buy. But the most challenging I see on general approach and mentality shift relying on AI solutions. That experience knows everybody who played chess on 8 bit computer in the mid 80´s. In a certain sense a computer is every time right. Today, equipped by big data, the AI solution could be recognized as more and more as definite. People generally, but public prosecutors particularly, are afraid of responsibility for decisions. The AI “big data advice” would be salvation for them in place of deeper human understanding of problem. At that moment it is only AI designer who can satisfy AI evaluator at best.
● How do you think AI will change your sector in the near future? Do you foresee disruptions?
Aside from processes, tools, etc. I see the major changes in the sector by introducing AI technologies generally. What position it will occupy in human life. It could have an enormous impact how we think about buildings, public space, transport, how we spend leisure time, who we will work etc. Outsets of autonomous cars could have technological claims impacting the shape of our shared streets. It would go to 3d when drone transport will become standard. The higher position and control we give to AI, the more adjustments of environment in the benefit of AI we are forced to sacrifice. For the architect's profession, the AI situation is not just about introducing new technologies in terms of designing, but more important I see needs for an overall evaluation on that phenomenon.
Pavel Martinek 20.3.2021