Amy Hodgetts discusses whether robots and technology render human workers obsolete
Technology has updated workplaces across all industries – and at such an alarming rate that many are questioning whether computers will take over the workforce. Currently, computer programs rely on human input, and because of this we cannot become completely reliant on technology. No matt er how sleek or agile these programs are, they will only create quality outputs if the input is correct.
Put simply, technology needs to be used correctly to produce correct results. Yes, an engineer can beneﬁ t from cutting-edge soft ware, but in the hands of a novice, it will not perform. In this scenario, there is really no replacement for human skill.
It is vital that computer-based assistance is treated as such. The successful link between computer programs and engineering skill varies depending on which part of the architecture, engineering and construction (AEC) industry they are being used in. To understand how this factor can impact their relationship, we must ﬁrst look at the three main stages of engineering design:
Concept design: At this stage, the majority of the design comes from the imagination of the engineer, supported by some simple sizing elements or calculations.
Drafting and analysis: This stage brings the concept design into the real world more earnestly, checking that it is feasible and how it will succeed. This stage is predominantly computer-based, using programs such as structural analysis soft ware to help engineers work to a greater degree of accuracy.
Detailed design: This stage is when, as the name suggests, the design becomes much more detailed. It is almost completely computer-based, with analysis happening in the background.
AI is no match for instinct
Anything that requires imagination and understanding will need human input. It’s not just the imaginative aspect that machines cannot replicate in full, though: ﬁne tuning, for example, still needs a guiding human hand in order to ensure the outputs are correct. While leaps and bounds are certainly being made in machine learning, whereby computers can make decisions based on historical data and records, it is highly unlikely that this will develop to the point where human skill and judgement become obsolete.
Human judgement is not infallible, of course. Mistakes can be made when writing programs to support design, or further along the line when inputting data into these programs. Either will result in an inaccurate output. For this reason, automated checking – whereby computer programs check the input against previous projects and their success or failure – has been a hot topic within the AEC industry. However, it is worth bearing in mind that the majority of engineering disasters have occurred due to something unusual – that is, something that has not happened in previous related projects. While rule-checkers help in situations where rules apply, they aren’t able to ﬂag something that hasn’t happened in previous records – i.e. something unusual.
The Millennium Bridge is a perfect example of this. The design met all the requirements of the design codes and yet it wobbled when it was walked on. Human understanding and ingenuity ﬁxed it.
As the world changes, engineers will make a value judgement to adapt their designs accordingly.
Both human and machine-based judgements follow formula in one way or another, and several structures and designs have had formulas developed exclusively for them. For example, the original formula for shell structures had to be created by mathematicians to ensure success. Now, with ﬁ nite element analysis, almost any form can be analysed — but that does not mean these forms are always sensible. There’s a certain amount of tension between architects and engineers surrounding this. Where engineers are seen as wanting functionality, architects are seen as wanting novelty ﬁrst. But this disparity makes for the perfect partnership – and the best designs.
See how structural engineering soft ware enables engineers to use automation to unleash their creativity at oasys-software.com