What is already possible with AI nowadays

For older people, it is one of the most ominous technologies in science fiction - for younger people, it is primarily a marketing buzzword: artificial intelligence (AI). But what can AI be for you? We set out on a search for clues...

AI has already become a part of everyday life

In 2018, Google presented a recording of a phone call at its I/O developer conference. It was about someone calling a hair salon to make an appointment.

So far, so unspectacular. Except that the caller was not human, but Google's AI assistant called Duplex. By now, at the latest, it was obvious: AI is here to stay.

Artificial intelligence will become so integrated into our everyday lives that it will be perceived less and less as something artificial. In some parts of daily life, it has already been present for years - perhaps not as effectively as Duplex, but already amazingly efficient.

Routine tasks that artificial intelligence has long since taken over include, for example:

• Machine translation
• Message generation
• Data mining and text mining
• Handwriting recognition
• Voice assistants
• Control of driverless cars
• Chatbots

If you look at the list, you will probably notice that some everyday tasks are missing - for example, purchase recommendations in online retail or route calculation in the navigation system. The reason: Behind such rather simple tasks there is usually "only" an algorithm, i.e. a clear action specification for the software.

If you order a bucket of paint and the shipping portal suggests a paintbrush, that's not a conclusion that would require artificial intelligence. Which leads to the question: What is the definition of AI?

Artificial intelligence: definition and delimitation

It was already mentioned at the beginning: Artificial intelligence has had an amazing career - from the horror scenario of many science fiction novels (in which artificial intelligence always includes human megalomania) to the all-purpose weapon for marketing, sales and software development.

This rise has led to many a piece of software being advertised with the AI seal of approval that does not meet the criteria at all. In addition, there is a lack of a clear, uniform definition. How can intelligence, and artificial intelligence at that, be measured and defined?

Especially since artificial intelligence does not have any intelligence of its own, but only makes intelligent decisions according to human definitions. Nevertheless, there are some points that can be agreed upon that distinguish "real" artificial intelligence from "normal" algorithms:

• AI consists of algorithms - but not vice versa.
• AI has the capacity for machine learning (ML), mostly based on artificial neural networks. These mimic the human brain.
• AI makes autonomous decisions based on given (as extensive as possible) data.

AI is therefore capable of learning - and can do so all the better the more data it is provided with. This information is either trained to it by humans or it collects it itself over time. Thus, with increasing experience, it can also become better and better. Depending on the degree of sophistication, this is also referred to as "deep learning".

By the way, all AI applications that you encounter today correspond to so-called weak AI. It is specialized in individual subject areas. Strong AI, on the other hand, would be equal to (or superior to) the intellect of humans in its entirety.

While weak AI is widespread, strong AI occurs only in thought models and in science fiction. For now.

Artificial intelligence: Impressive, not scary

According to Statista, the revenue from enterprise applications in the field of AI is expected to reach around 31.2 billion US dollars by 2025 - a figure that illustrates the impact of artificial intelligence on the economy. But AI or rather AI impresses even more with its performance - let's take a look at exemplary current achievements:

• Thanks to AI, image quality is improved in low-bandwidth video conferences - AI calculates facial movements instead of sending moving images. (Link)
• An AI calculated the blend (mixture) for a Swedish whisky. (Link)
• AI helps conservationists track populations of endangered species. (Link)
• Thanks to AI, a novel prosthetic hand can learn independently and be controlled by thought. (Link)
• AI improves tumor diagnostics. (Link)

Artificial Intelligence in Can Do

Let's summarize again: Artificial intelligence, when provided with the appropriate data, can make decisions on its own (and faster than humans). Doesn't that sound like exactly the kind of tool you could use in your day-to-day project management work?

The PM software Can Do uses artificial intelligence where it can play to its strengths particularly well: in resource management. The AI structures and condenses the volumes of data that contribute to resource planning decision-making, weighs the availability and capabilities of all resources in the company, and deploys them in a way that best supports your project.

And it does so at lightning speed. As befits a true AI, Can Do is also constantly learning. Artificial intelligence gets better with every challenge!

Conclusion

The question from the beginning, whether AI is an opportunity or a bugbear, doesn't really arise anymore these days: Too often, "weak" artificial intelligence has already proven its great benefits.

Be it in science and research, in everyday applications or in software optimization - as in the case of Can Do. Whether "strong" intelligence has more of a potential to become dangerous to "humans" is nothing more than dreams of the future, speculation and fantasy. And until then, AI makes life - and your project management work - easier for all of us!

Sources & Inspiration: spiegel.de, wikipedia, golem.de, microsoft.de, wikipedia, statista.de