Reading time: about 4½ minutes – read, like and comment.

In 2019, Artificial Intelligence (AI) was the hot topic of the scientific year. A ship travelled the length and breadth of Germany, allowing people of all ages to engage in a stimulating way with the idea. However, there was more than just publicly-funded scientific communication on the AI agenda. Various book publications, newspaper packs, and TV and radio features elevated AI into the consciousness of those who had not already come acr  oss it in blogs and through other digital channels. Views on the potential of AI ranged from respectful subservience to disparagement. “Think of Artificial Intelligence as a three-year old child”, wrote Spiegel on 3rd December 2019 (link is external). At the dawn of a new year, is there anything still noteworthy about this topic for Adult Education? After all, AI tends to be associated more with robots and self-driving cars than with educational events.

Use of AI in Adult Education

On closer inspection, the technological developments also affect adult teaching and learning. Not that it is anticipated that course leaders and programme planners will be replaced by machines in a few short years. Ongoing investigation into the potential for substitution on the labour market currently envisages, particularly in the case of professions where there is a high level of non-standardised guidance, that the human workforce will continue to remain important. No, there is no risk that humans will become superfluous in Adult Education; the tools surrounding and supporting them will merely become increasingly infiltrated by AI.

Let’s imagine a person by the name of Roland in the year 2025, who wishes to attend a further education course on the topic of first aid.  Unless the programme brochure of his trusted educational establishment is enough to attract him, Roland will avail of more or less intelligent digital aids, during an internet search, to research a course. The offers he is shown and has recommended to him will depend to no small degree on his surfing behaviour to date and on his preferences in terms of location and value. Whether “only” defined algorithms are at work in this case or whether AI is already involved depends on the learning ability of the algorithm. By 2025, search instruments will be even smarter than they are today as the result of active self-learning. Probably, although that remains to be seen, the recommendations will be based on the principles of similarity and not on the principles of productive opposites.

Future scenario 2025

Assuming Roland decides on a blended learning course, he can then expect to encounter AI both in the classroom-based as well as the online components. Let’s start with the first classroom-based session, during which Roland is frightened by the AI-controlled doll, which starts coughing after the first correctly performed resuscitation attempts and subsequently thanks him warmly. The lecturer, Doris, has brought teaching materials with her that were selected with the help of algorithms in exactly the same way as Roland’s course search. Doris may even have availed of a tool that discloses the computer-based search quality (such as Kansas for language learning courses).

Elements of AI might be more pronounced in the online aspect of his course than in the classroom-based part, however stakeholders, who want to exploit the potential of AI, are more frequently involved in the development of electronic learning courses. Roland very quickly makes friends with his online tutor, who – without becoming grumpy – is even there for him with advice and help after 11.30 p.m. Roland only became aware of the fact that there is a chatbot behind the tutor when he once posted a witty comment without the relevant winking-face emoji and that was the reason the tutor didn’t understand the joke. It might still be a pipe dream in 2025 for AI to actually have a sense of humour and for this quality to be such a standard feature that learning software providers work with it.

The potential of learning analytics, which will underpin learning software, is not so much technically complex as morally and legally questionable. Roland’s learning behaviour will, namely, be blithely recorded (tracked), insofar as that is legally permissible and Roland has agreed to it: his learning times, his learning progress, his mistakes, and his interactive behaviour. From this, the software is able to generate personalised feedback, which he and Doris will use to discuss the next appropriate learning steps during the next classroom-based session. For Doris, this means that her workload becomes somewhat lighter, while in Roland’s case a vague feeling, somewhere between unease and incredulity, sets in.

Even the learning system has continued to learn in the meantime. It recommends offers to Roland during his next online session that the algorithm considers particularly suitable. Intelligent recommending is one of the main areas in which machine learning is used behind learning software. The majority of recommendation systems are probably still trained in collaborative or content-related similarity. This means that the system checks which interactive behaviour with the system or which thematic preferences Roland used for learning successfully. This will then be intensified in each case. By 2025, progress might have been achieved vis-à-vis this kind of somewhat primitive paedagogic intervention.

AI as the subject of Adult Education

So much for a scenario that shows the areas in which AI has or is due to become part of Adult Education activities.  However, a second level is at least as significant: AI as the content or subject of Adult Education.

In light of the astonishing (useful just as much as threatening) potential of AI, adults should have basic AI competence in order both to be able to participate in an AI-supported living and working environment as well as to counter a potentially dangerous AI reality with sound philosophical arguments.

The adult educator, Michel Brendel, has written a book entitled “Future Intelligence”, which has one objective above all others: to issue a call for widespread discussion across society of AI. This role would be tailor-made for Adult Education provided it became capable of speaking out on the topic itself, so as to be able to take part competently in programme planning and orchestrate the dialogue.

All of this initially requires becoming familiar with the topic. Through the free online course developed in Finland “Elements of AI”, for example. The data privacy statement reveals on which of your steps the system will be reading along with you. Because it will learn itself.

About the author: Dr Peter Brandt is head of the "Knowledge Transfer” department at the Deutsches Institut für Erwachsenenbildung [German Institute for Adult Education] - Leibniz-Zentrum für Lebenslanges Lernen e.V. [Leibniz Centre for Lifelong Learning] in Bonn.