These days artificial intelligence (AI) is on everyone’s lips. Some conjure up the end of (human) work, others are afraid that Skynet will soon become a new reality, and some are convinced that AI will solve almost all human problems shortly.
First of all, it can be said that all of these scenarios are highly unlikely. Few technologies have been hyped as much as artificial intelligence. The images in our heads are accordingly, in many cases, very far from what the current reality looks like.
Nonetheless, only a few technologies have such a fundamental impact on our lives as AI systems are already doing today. Both for good and evil, the potential is insanely great, and development is proceeding at a breakneck pace. It just simply cannot keep up with the images drawn in minds, films, and books.
AI: What is it about?
It is utterly impossible to describe the technological and social complexity of this technology in a short article. But fundamentally, we should have at least a rough understanding of three key concepts: Artificial Intelligence ( Artificial Intelligence ), Machine Learning ( Machine Learning ), and Deep Learning.
The legendary artificial intelligence describes all systems (machine nature) with a particular form of “intelligence.” Such systems, like humans, can solve problems that require some learning and understanding. A very rough distinction is made here between strong and weak artificial intelligence. While a strong AI has general intelligence and can solve virtually “all” problems, a weak AI is limited to individual problems and application domains, for example, speech recognition.
In reality, at the moment, there are very many weak but no (known) solid artificial intelligence.
Machine learning is a sub-discipline of artificial intelligence. Machine learning encompasses various methods and algorithms that allow computers to learn information from data and make decisions based on this.
Two of the most common tasks in this area are classifications (spam or not spam?) And predictions (how will the price change in the future?).
The procedures that are used here are relatively easy to understand. In most cases, the algorithms involved are mathematically easy to understand.
A distinction is also made between supervised and unsupervised learning. In the first case, the computer receives “learning material,” for example, a series of examples of what a spam message could look like. In the unsupervised case, the system should draw information from the data without human intervention and material (training data).
Deep Learning is now one more sub-discipline of machine learning. However, instead of using relatively traditional algorithms, the aim here is to model the functioning of human brains.
For this, (deep) networks, which consist of many levels, which consist of artificial neurons, are simulated. The data then flows through these networks, similar to the way they do in the brain. In the meantime, there are incredibly complex network architectures that can independently generate completely new insights from the data.
WHAT DOES THAT BRING US?
Various methods from the field of artificial intelligence are already being used in almost all areas. In principle, we can distinguish between three cases :
- The AI replaces a person and takes him or her work. An example would be a self-driving car or a robot that takes the strain off people.
- AI does something that a human could do in principle, but much faster and on a larger scale. This is always exciting when the amount of data is vast. People can, of course, distinguish spam from important email, but there aren’t enough people to handle the large volume of data in a meaningful way.
- The AI does something that a human or humanity cannot. We already see cases, for example, in medicine, in which AIs have found patterns in the data that have remained hidden from scientists.
In short, in the future, we will always rely on AI when the task is either too “boring” or too “big” (big data). Of course, it is also very likely that better and better AI systems will solve problems in the future that we don’t even think about today.
On the other hand, many critics are concerned that AIs could become a real threat to humanity. Here are four of the top concerns and issues to keep in mind:
- What happens if a “bad” person is the only or first person to access a mighty AI?
- What happens when AI systems, which may also have decision-making power, learn the “wrong” things that are considered morally unacceptable?
- What happens when AI systems drive the social divide to the extreme because those who have access to AI have an unlikely advantage over those who are “only” human?
- What happens when we as humans can no longer understand which decisions an AI made for us and why?
To get these problems under control, we have to be proactive in various places. We need to make sure that we all have a basic understanding of these technologies. We need to make sure that research and development are done publicly, and that technology is democratized. We have to make sure that we develop moral, ethical, and legal norms that serve as guiding principles.
WHERE ARE WE STANDING?
As already described above, we are still a long way from a general, vital artificial intelligence that equals humans. But we already have effective AI systems that make many fundamental decisions for us.
Whether we get recommendations on Netflix or Spotify, scroll through Instagram or somehow see advertisements – there is almost always an AI in the game that learns from our behavior what we want or should see next.
Many essential technologies, for example, from the area of machine learning, have meanwhile been researched and worked on so well that their use is relatively easy and inexpensive.
An exciting example is IBM’s Watson; an AI that, in addition to many other tasks, supports doctors in over 230 hospitals in deciding which cancer therapy is best for which patients.
Another example comes from the financial industry. Almost all large banks rely on AI systems to monitor transactions and keep an eye on the flood of banking transactions. On the other hand, sit traders who try to use AIs to make profits on the stock market.
The list of examples could be continued as desired. The only clear thing is that AI is already being used in almost all products at some point. If we don’t act completely stupid, that shouldn’t be a problem in the future, but rather a blessing!