We would like to clarify a few points upfront:
- Most AI applications today, and in the very near future, apply to narrow and specific areas like recognizing objects in an image, speech recognition and text translation. In contrast, human intelligence applies to a very vast area. AI technology will take several decades to achieve the breadth and depth of human intelligence.
- AI today is not perfect and cannot deliver 100% error-free outputs. As we all know, nothing is error free, not even human predictions and decisions. However, the performance of (narrow) AI systems is superior and more accurate than that of us normal people, making it attractive in many applications. In the initial phase of deployment, AI decisions will be applied as expert opinions or guidance. Final decision to take action can still be taken under human supervision following the AI recommendation. At some point in the future, when a high trust in AI decision is established, the human approval will most likely be seen as redundant and dropped.
- In assessing errors and mistakes in performance, we are much more tolerant to mistakes made by humans, caused by forgetfulness, being tired, being distracted, falling asleep etc. However, we expect AI machines to be perfect and error-free. Eventually we will recognize that machines make far fewer mistakes than an average human, are cheaper, and easier to maintain. That will tip the balance in favor of AI solutions.
- Every technology can be used and abused. AI delivers significant benefits and innovations for almost all industries. However, we are aware that a powerful technology like AI can also be used to do harm. Science fiction often paints many negative scenarios of how AI will dominate the world, exploiting and subjugating humans to menial and degrading roles. Much more concerning is the imminent issue of job displacement, retraining, and unemployment of those who lose jobs to AI and robots in the very near future. As a consequence and number of other problems can arise, like growing inequality, social tensions, and geopolitical shifts. Unfortunately, at the time of writing, we have not come across any satisfying solutions for these issues.
The first section of this book focuses on making AI understandable, providing an easy to understand background on some fundamental questions: What is AI? What does it really do? How does it work? How did it evolve? How can machines be made intelligent? How do they learn? How does it fit with other digital technologies? What makes it useful in business?
The second section is focused on putting AI into action for your business – in an evolutionary as well as in a disruptive fashion. By using a few case studies we illustrate how AI is being used to create customer and business value. What skills and competencies are needed, how to develop or acquire them, and how to integrate AI into a product or process? We explore some issues like the role of humans in the future of collaborative work with AI, the future of work in the society, and the social consequences of the AI revolution. Lastly, we seek to address the outlook – what is the end game and where are we heading?