Artificial Intelligence is a term that nowadays popular culture has loaded with exceedingly varied and colorful meanings. Indeed, the term comes to us with its own very personal semantic baggage. From Asimov to Westworld, machines that act and think like human beings still remain a mainstay of science fiction.
The fact of the matter is that the numerous attempts to employ AI (Artificial Intelligence) in various types of activities are today pointing out its not insignificant limitations. Why? Machines do not make good decisions, and they are not creative. Examples of these limitations have not been slow to manifest themselves after their first uses: think, for example, of the damage suffered by Tay, Microsoft’s Twitter bot, at the hands of Trolls. Tay offline, without going through the gateway. During the past few months, the AI has been trying to write the first sentence of a novel, failing resoundingly.
Examples supporting how AI is not sufficient for the purpose – as it has been assumed and used so far – are therefore commonplace. That is why a new definition is needed: AI, yes, but “Augmented” instead of “Artificial.” The new definition proceeds properly from the experiments carried out. It makes well-founded sense today – on analysis and studies of various kinds – to speak of Augmented Intelligence, that is, of a system of AI technologies strictly in support of humans, not in substitution.
Augmented intelligence does not attempt to imitate the human mind; on the contrary: it amplifies it, supports it. In two words, it makes it better. Each of you has likely already come into contact with Alexa, Siri and OK Google: all AI technologies. Therefore, you will certainly all know that these technologies are not capable of making any decisions on their own, in the absence of human action. Instead, what they can without less do is to improve the ability of humans to make good decisions by serving as a direct bridge to society’s collective memory: the Internet.
To sum up: AI can take on exponential and extremely profitable value.
Augmented Intelligence metrics
AI (Augmented) is able to connect purely human qualities and skills (creativity, learning and decision making) with qualities and skills that are the exclusive preserve of the Machine (data analysis and data recording). The insight goes without saying: in this new playing field, both humans and machines come out appreciably improved. Synergy thus appears to be the only way forward, in order to make this much-discussed meeting of two worlds pay off to the top (thus abandoning the unsuccessful assumption that one must surpass the other, in the false war posed in the literature on the assumption).
Researchers today are focused precisely on Augmented Intelligence, rather than Artificial Intelligence. According to the authoritative opinion of Andrew Moore, dean of Carnegie Mellon’s School of Computer Science, 98% of researchers today are working on creating intelligent systems that help people make better decisions, rather than on creating machines that can make them on their own.
This, then, is the trend in AI-related technology. That is why it is no longer so anomalous to talk about Augmented Learning.
It sounds strange, yet each of you has already used this technology. For example, each of you will have used a calculator to perform mathematical operations. Well, you have used a type of “enhanced machine learning,” aka: Augmented Learning. So how exactly does Augmented Intelligence combine – and according to what parameters – with e-learning?
The rise of chatbots
Chatbots are probably the best example of AI in L&D (Learning & Development); a kind of Alexa of e-learning. Should a student hypothetically need support for a given learning activity, an e-learning chatbot could effectively guide – and track – him or her toward that learning path that he or she would not have automatically considered through his or her own means.
A chatbot may also be able to act as a virtual assistant for the student within a training pathway; if a student has not completed a course as planned in the instructional design, the bot can remind him or her to proceed, as well as any deadlines for doing so. This type of function always puts the student in a position to be able to complete their training path correctly, and save them from forgetting a preparatory course. The Training Manager, on the other hand, comes out safe from having to remind his team.
These “reminders” via chatbots are also functional in saving useful information. A chatbot can enhance the skill acquired by your learners in a given training course by acting as a reminder precisely in relation to the practical use of that given skill. Learners have a tendency to forget very quickly what they learn: Edgar Dale’s Cone of Experience explains in depth how people retain only 10% of what they read and 20% of what they hear. By practicing and enhancing the skill learned, however, can concretely grow that value to 90%. A chatbot can help. For example, if an employee takes an evening training course from home, one potential and effective use of an e-learning chatbot is to remind the student of what he or she learned the night before, just as he or she is at work. In this way, the employee-learner will be in a position to readily put into practice what he or she has learned, settling the notions as best he or she can.
Another effective form of AI applied to elearning is personalized training. This task is sometimes effectively performed by the LMS platform, which is inherently already programmed to change courses based on the individual student’s performance. Sometimes it is therefore beyond simple to provide students who fail a given test with additional material that they can use immediately.
Other times, however, it is necessary to implement a form of personalized learning that can be carried out through chatbots. Indeed, the colloquial nature of the chatbot makes it a perfect private tutor, able to answer questions and meet the individual needs of each learner.
Chatbots can also be used to provide thematic responses in a manner untethered from a specific training course to anyone in the company who requests it. Suppose a company requires all new hires to take a course on internal procedures and policies: the course will certainly be updated regularly – meaning that new hires will always receive the latest information – but longtime employees may not be fully educated about the updates instead. Therefore, a chatbot is ideal for providing information about updated procedures on a target basis (i.e., to those students who have passed the onboarding level), so that the entire “classroom” is on the same page.
Using AI to collect and analyze data on students
LMS platforms are generally able to tell us a lot about our students.
This kind of data is of enormous value to CLOs – Chief Learning Officers (Corporate Training and Development Managers), who obviously want to know how students are behaving and how much they are learning, but it is obviously impossible for the human mind to record and manage this amount of data simultaneously.
Augmented intelligence can concretely and effectively support trainers and L&D Managers precisely on data collecting and data analysis. Such a model already exists in the health care sector in the States. Indeed, as you may know, physicians have access to an enormous amount of information – as well as diagnostic images – about their patients.
The system-education is also beginning to approach this kind of model, implementing, for example, “alert systems” for those students enrolled in courses that are online and blended in nature. These systems are able to monitor student behavior in and out of the classroom, recording any dropouts both online and offline. As a result, when a struggling student is identified, for example, the program simultaneously alerts faculty, tutors and the student. Ergo, this can no less become a winning model for Training & Development.
AI as a tool for e-learning
When it comes to e-learning, one must be well aware that tools represent the basis of online learning, at least as much as the content of a course.
This is exactly what augmented intelligence represents in the L&D field: an excellent training tool, employable as an integral part of a course, as a learner support, as an enhancement tool for a given training program, or as a stand-alone tool suitable for meeting learner needs. It effectively supports trainers in analyzing data, identifies at-risk students, and contextually offers individual mentoring.
In conclusion, we do not know the extent of what augmented intelligence can concretely come to do in e-learning. As with all digital technologies, it will take (as usual…) human imagination to track AI to its full potential.
Thanks to A.J. O’Connell – Litmos’ witty pen – for the valuable contribution.