Updated: Mar 15, 2018
Artificial Intelligence (more commonly referred to as AI) is a hot topic at the moment and why wouldn't it be, it really does have the power to change the world as we know it. We could discuss this for days so I'm going to try and focus on it's relevance and applications towards the engineering industry. In this blog, I'm not going to look too much at the numbers, we are simply going to analyse the current capabilities of AI and some control methods that we would recommend. If you have differing opinions, please discuss them with us in the comments section below.
AI is a very hard topic to define, especially for engineers as it is just a natural progression from software engineering applied to machines and systems. In truth, it is a machine and / or systems ability to display human-like skills such as problem solving however to many engineers this has just been seen as automation. The difference lies in the way that AI is able to learn allowing machines to become even more capable at performing tasks. This has been amplified by companies such as Boston Dynamics (see one of their latest creations in the video below), looking at AI from a robotics perspective and DeepMind, from a systems engineering perceptive. Even so, it is still very difficult to differentiate between automation and AI for most engineers.
Atlas, Boston Dynamics, Nov 2017
Whatever we call it, AI has become so capable that many tasks can be performed more effectively by a robot or autonomous system than a human. This in turn creates a major problem for many companies and larger societies. If a robot can outperform a human, why would we use a human to perform this task. Most examples of this come in middle of the range jobs such as fabrication, warehouse work and civil servant jobs. We can see this through the automotive industry implementing a nearly autonomous production line, packaging warehouses utilising robots and machines and even the reduction in checkout staff through the implementation of self-service checkouts. Plenty more examples can be given but the key point when looking at a business case is to look at the cost of implementing such a system versus the cost of maintaining the current strategy.
Ford Model T Production Line vs Modern Toyota Body-in-White Fabrication on the Production Line
Focusing back on an engineering perceptive, have autonomous-based systems and AI taken away from any jobs? Yes, there are far less workers on the production line but the amount of people to implement and maintain these systems has also increased so the total number has remained roughly the same in cases where businesses have maintained their size / revenue. A prime example comes in a previous blog discussing "industry experts". I imagine many "engineers" performing mediocre tasks on the production line were not robust enough to adapt to the new working conditions. For myself, as a Design Engineer, I can already see the industry changing with the use of highly advanced parametric modelling (also referred to by others as AI), however I am preparing myself to ensure that I have the capabilities to develop and maintain such systems. A machine may well be able to create a turbine engine based on a set of parameters but in no way will that model be realised without the use of checkers (highly experienced engineers) or be reused without the system being continuously maintained. In terms of engineering, widening your skill sets as much as possible will only increase your chances of being employable when AI systems are implemented. You may be tasked with helping develop, maintain, check or identify such systems. Focusing and deeply understanding the core skills that make up engineering is the best way to widen your skill sets and prevent autonomy from taking your role away.
Looking at AI from a purely deep learning perspective, I believe the technology is far greater than those outside the industry realise. It has the power to analyse millions of different scenarios within seconds to determine the outcome. If we use these systems for good, it can only improve the current technology, safety and economy. However, in the wrong hands it can create many issues such as advanced warfare, security and inhumane technologies. Herein lies the fundamental issues with the obvious answer being regulation, but how do you regulate something in when you don't know its capabilities? Additionally, international regulation is very difficult. An example of regulatory errors comes from the aviation and aerospace industry. It took many plane crashes before strict FAA and EASA (and many more) regulations were put in place. I like to think that some governing body is looking at this as we are on the brink of huge technological excellence if we can apply such technologies to everyday society. AI shouldn't, but certainly is, something to be wary of. The potential advances includes improved human cognition and memory, health services improvements, economical stability (analysing the markets before decisions are made) and limiting human error (for example in driver-less vehicles) just to name a few.
To conclude, the answer is AI simply isn't taking away jobs. It is an individuals capabilities that prevent or terminate their employment. My recommendations, as in previous blogs is for engineers to be able to adapt to new technologies and have a wide array of skills which can be applied to many different engineering problems. AI certainly has the power, and will in my opinion, change the world. Whether that be for good or bad is yet to be seen but I do hope that those in power listen to the experts and regulate these technologies before they can be used in ways in which we cannot yet comprehend.
I hope my opinions are shared but if not please discuss below. Additionally, if I have missed anything out, please also add that in below. I will also be setting up an AI forum on this site in which members can discuss new advances and look at the latest news.