Demystifying Artificial Intelligence
In this article we want to introduce you to AI and explain the most important facts and most importantly: why is adoptr.ai called adoptr.AI?
What does AI even mean?
AI is a hyped and often overinterpreted word, which often goes hand in hand with a subtle anxiety. The first time I came in contact with AI was watching the movie “I, Robot”, which, to be honest, did end in anxiety and unknown about the potential outcomes. Thanks to the dystopic scenarios created by the entertainment industry, there often is a quite palpable sceptic. And that is exactly why AI needs to be demystified! If you look at the term unemotionally the definition could be phrased as the following: as a part of computer sciences, the main goal of AI is the automization of intelligent behaviour
What does the AI stand for in your company’s name “adoptr.ai”?
We have decided, to offer exactly the tools needed to help our customers tackle the step in their digital journey, which causes them trouble: the change contributed by technological innovations. As early adopters of just those innovative technologies, we feel the need to be open and know exactly what the new technology can and cannot do in order to guide our customers and partners through the adoption process. But behold, we are more than just adoptrs!
We have knowledge and experience as thinkrs, enablrs and changrs, and this experience allows us to assist customers and clients as Innovation and change managers and to help boost their businesses. We see the benefit of AI and already use them in our daily lives, such as in our work time tracking tool, which is assisted and made smart by Machine Learning.
What is the difference between Machine Learning and Neuronal Networks?
Easily put, in both cases the main goal is for a computer to solve a given task in a self-sufficient manner. The complexity of the computational model however is crucial. In the area of machine learning, algorithms are fed with data and they then notice a certain pattern. This learning process can be steered, for example by using pictures, in which dogs and cats are to be differentiated. If with every picture either the solution “dog” or “cat” is supplied, the algorithm will only see dogs or cats. That area is called supervised learning.
In unsupervised learning however, the algorithm would search for a solution pattern and an outcome could be, that the differentiation would not be made between cats and dogs, but between differently coloured animals.
Another area of machine learning is the so-called reinforcement learning. In this case the algorithm is steered by incentives. So basically, depending on the decision of the algorithm it is either praised or punished – like the training of dogs!
How successful these models are, relies on the amount of data provided and the amount of data the model can process. This is where the massive benefits of neuronal networks come together. Thanks to its structure, a neuronal network is capable to notice connections, that remain hidden to simple machine learning. This is due to the fact, that it is built in several layers, which can process information more efficient and effective. However, intricate the setting is or how many layers are used, it is called a deep learning network.
Why can AI be helpful?
This can be explained by the magic triangle: the combination and dependency on the three corners: quality, cost and time. This indicates, that in order to improve one or two corners, the other corners will decrease. If you want to produce something faster and in a higher quality, the prices will rise.
With AI you can gain improvements in all three areas. Our automated time sheet is a great example: due to the fact, that the machine learns from patterns, on which project was worked for which amount of time, it saves us the hassle of keying in everything manually. This way we don’t only save time, but also money and our productivity and satisfaction are improved.
Can AI be dangerous?
You should be aware, that in the area of AI, especially the end-user, works with a “black-box”. So basically, the end-user is unable to see the complexity of the algorithms behind the AI, which is not expected of him in any case. But it is important to note, that especially in sensible areas, algorithms should not be trusted blindly. The CEO of Microsoft Germany leads with her statement: “What improves the combined work of mankind and machines?”. AI should be viewed as assistance, but it remains our decision, where it is and where it is not used.
Which tangible use cases of AI are available today?
AI can already be found in a lot of different industries with many use cases. One easy example that most of us know of, is that our mobile device is able to unlock the screen, simply by scanning our face. We trained the system to recognize our face from any other face.
In the business sector the use of AI will become undisputable as soon as a large amount of data is analysed, and patterns are recognized. Especially for small and medium sized companies heaps of use cases will arise, when more IoT-based business models are established: the amount of data, that is generated by sensors and products that are connected to the internet, can be processed in a more efficient way with AI.
The potentials of AI are undisputable, and our goal is to view the whole, the big picture: we want to raise awareness, we want to be aware of the anxiety and misunderstandings and our aim is a successful and long-standing partnership that brings benefits to our customers and partners. Feel free to contact us!