Digitalisation and the new technological possibilities that artificial intelligence (AI) brings are driving the biggest social and economic changes since the industrial revolution. These are associated with opportunities and risks. Without the right political, economic and ethical framework conditions there is a risk of uncontrolled development and a negative impact of AI.
AI ultimately affects all industries more or less heavily. AI or specific forms of it, such as machine learning (ML), can be used in a wide variety of application scenarios. First and foremost, they will be increasingly found in areas where large amounts of data need to be analysed and evaluated. AI systems are better equipped than humans for analysing massive amounts of data, correlating the data with various reference points and thus to create a better base for decision making.
AI is shaping the way we work
AI has already entered many areas of our working and private lives. Already classic examples of this are Alexa or Siri. By means of smart algorithms, machines today are capable of doing incredible things with facial and speech recognition. With error rates of under five percent, many systems can perform better than humans. In image recognition, which is also used on Facebook or in self-driving cars, computers are now far superior to humans. And online retailers or search engines, for example, use ML to optimise the user experience and to create buying recommendations. In short, AI and ML are already fixed – and also largely accepted – components of our day to day lives.
The impact of AI, when it comes to our working environments, AI, ML and digitalisation bring a significant sea change, as they alter or expand the human activity spectrum. This means employees’ work profiles and what is required of them will change considerably. A typical example is industries where more and more machines and robots are used, such as the manufacturing industry.
AI brings risks
With regard to the use of AI, there are two principal risk categories: on the one hand, there are risks connected with society and humans and, on the other hand, there is a risk of dependence on technology. Many are concerned about machines and AI taking over human activities. This is not just about apocalyptic fears and Terminator-like scenarios, which have also been cited repeatedly by Tesla founder Elon Musk, but about more elementary, existential fears. People ask themselves questions such as: How do I fit into the digital future when intelligent robots take over my job? Do I still have the right skills? The older generation especially is very sceptical about the technological development and increasing use of smart machines.
However, it should be remembered that the world of work will definitely change. Jobs that are indispensable today may not be relevant in a few decades’ time. Currently, the impact of AI is seeing a great need for data scientists and developers, but in a few years from now, machines might be able to do these jobs better and faster. The technical development and the concerns of society make it imperative to start socio-political initiatives that develop and implement not only the technological opportunities but also educational and digital policies. Without a concept, Britain will weaken its international competitiveness. The second risk category concerns the question “What decision-making powers do we give machines and AI?” There is a risk of becoming too dependent and unable to reproduce how a machine or algorithm has come to a conclusion. Neural networks, which have millions of connections and interactions, make decisions that they cannot adequately explain due to a lack of communication skills and the complexity of their “thought processes”. Machines are not emphatic narrators, they know more than they tell us, and they do not understand our information needs.
In any case, in some critical areas and rapid decision-making processes we will eventually have to rely on information from machines – take for example the decision that the immediate shutdown of a nuclear power plant is necessary, since the complex trains of thought involved in such events cannot be traced and understood fast enough by humans. It is also often overlooked that even machines can have prejudices based on information from the past. An example would be a recruiting system that supports the selection of optimal candidates by analyzing thousands of historical decisions made by the company. If a recruiter in the past – for whatever reason – preferred men over women, then this “prejudice” is also inherited by the machine and becomes part of the algorithm. In addition, there are already AI applications today that develop their own successors without human input, as in the example of Google’s AutoML project. The impact of AI could mean it creates a dynamic that can no longer be controlled, which is indeed disturbing. And, it stresses the urgent need for a moral and ethical framework.
Impact of AI and cybersecurity
Before any use of AI technology, elementary questions must be answered. The area of cybersecurity, which uses AI and ML for better threat recognition and the prevention of financial or operational damage caused by cybercrime, is a good example. Threat Detection and Threat Intelligence are the keywords here. The following questions must be considered before AI is applied:
• What do I want to achieve?
• Which ML methods best support my goals (such as supervised learning, unsupervised learning, decision trees, or deep learning)
• What impact does AI / ML have on my organization and employees?
• What risks might be created by the changes?
• How much decision-making power do I allow the machine?
• Where and when do we expect human intervention?
Overall, industry and society are still at the beginning of the change process brought about by AI and ML. New opportunities will drive innovation and create new ideas. We will see mistakes, but also unexpected successes. However, one thing must not be overlooked: The effects of AI are serious. Humanity, according to the World Economic Forum, faces the fourth industrial revolution – with social, economic and societal implications. Therefore, politics and business must work together – here and now – to create the appropriate framework conditions that will set the course for future generations.
Kai Grunwitz was appointed Senior Vice President EMEA at NTT Security in July 2017. He is responsible for both the business and the development of go-to-market strategies in the consulting and managed security areas of NTT Security. Kai has over 20 years of experience in top management positions in the IT industry.