Meta navigation

Artificial intelligence

Artificial intelligence

Artificial intelligence is networking

AI-based systems and networks are becoming established in our digitalised everyday life. Everything is now smart, from toys to the household and the even the company-wide “central brain”.

Beat Hochuli

Artificial intelligence (AI) is currently experiencing a veritable renaissance – or perhaps even its actual birth. After the initial euphoria that accompanied it in the 1970s and 1980s, for a long time it tended to be ridiculed and dismissed as the fantasy of a few freaks and nerds – though not in Hollywood, of course, whose apocalyptic story outlines only reinforced the impression of an unrealistic, illusory world.

In the maze of intelligence

In the age of big data, the Internet of Things and self-learning algorithms, this appraisal is now being completely revised. At almost every event dealing in the topic of digitalisation, the merits of today’s, and above all tomorrow’s, smart systems and networks are being sung. This unavoidably raises the questions of what exactly AI means. The answer is as easy as it is – for many – unsatisfactory: A crystal clear, unambiguous definition does not exist. Hans Magnus Enzensberger, in his judicious and amusing essay “Im Irrgarten der Intelligenz – Ein Idiotenführer” [In the Maze of Artificial Intelligence – A Guide for Dummies] analyses the innumerable facets of natural, i.e. human intelligence, and when it comes to artificial intelligence, all that is left to say is: It exists, and it presents itself in all sorts of different guises. One thing is, however, certain: In an age of all-encompassing digitalisation, it is increasingly manifesting itself as networked AI.

From AI Trump to the nursery

The artificially intelligent Donald Trump, which was created by Bradley Hayes from the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (MIT) was meant more as humour than anything else, but nevertheless has a serious basis. The bot is called @DeepDrumpf, was fed with data from Trump’s speeches and contributions to debates, and uses this material to create tweets of its own from scratch. In doing so, it proceeds letter by letter, meaning it doesn’t just piece fragments of existing Trump statements together. However, the tweets that @DeepDrumpf fabricates still are astoundingly similar to the tweets sent by the original – even including a good portion of nonsense.

Screenshot Twitter-Account @DeepDrumpf.

The method does, however, become serious – in the nursery, of all places. That’s right, this of all places is what has been singled out as an arena for AI by the developers of intelligent toys. The talking dinosaur made by Cognitoys, for example, is based on the processing power of IBM’s supercomputer Watson. The artificial reptile understands what the child is saying, and is capable of learning more and more about its real-life playmate of its own accord. According to its manufacturer, the dinosaur learns “what the child’s favourite colour is, favourite toys and other things the child is interested in, using this knowledge to cultivate an individual relationship. And more yet – the toy has a personality of its own, which changes over time”. Welcome to the nursery of the future!

The CongiToys Dino learns about a child’s interests.

AI-based living and working

Robert Bosch Smart Home GmbH, a recently founded subsidiary of the manufacturer of household appliances, Bosch, has set its sights on the entire living area. It is developing solutions for networking all intelligent appliances in the home, and for managing these appliances using an integrated app, that also understands appliances made by other manufacturers which have a compatible network.

“80 percent of the knowledge that large companies have is not written down, but rather is stored in the minds of the employees.”

Leveraging everyday business to be more intelligent is what the Zurich-based start-up Starmind, with its central brains based on AI made for companies, is aiming to do. According to company founder Pascal Kaufmann, this is essentially about networking all employees to ensure that the shared knowledge existing in a company can be optimally used for any problems arising and the solutions they require. When the artificial brain is asked a question, it searches for an answer by passing the question on to all employees, at the same time as realising its own learning process. “Eighty percent of the knowledge that large companies have is not written down, but rather is stored in the minds of the employees – who learn by experience and intuition,” as Kaufmann argues the main reason for employing Starmind. Swisscom has also been using the self-learning Starmind networks since 2013. The “Ask the Brain” system makes knowledge from the company's network available to every employee by automated means.

Of course, data protection and security concerns also need to be consistently factored in when it comes to all these AI-based developments. This will not, however, present an obstacle to the continued boom in networked AI.

Beat Hochuli

is a freelance ICT journalist and lives in Kota Kinabalu, Malaysia.

More on the topic