Artificial Intelligence and E-learning: where do we stand?

It has been 70 years since Alan Turing (mathematician, cryptographer, philosopher as well as one of the fathers of computer science) introduced theories for building "intelligent machines" to the scientific world. Today, many sectors are looking to Artificial Intelligence as a hotbed of possibilities. Referring to e-learning what is the relationship? What are the possible benefits and future developments?

share on

Reading Time: 15 minutes

Summary

What is meant by Artificial Intelligence.

Artificial intelligence is not a science-fiction movie supercomputer or a super-powerful software that can replicate the thought patterns of a human being.

Rather, it is the set of scientific and engineering disciplines that gravitate and operate around the concept of the “intelligent machine.”

Artificial intelligence in daily life

In practical terms, what does artificial intelligence consist of?

It is now talked about on a daily basis, its characteristics are increasingly explored, and new horizons are imagined. There is one thing, however, that many people do not dwell on: we are dealing with artificial intelligence much more than we expect and for much longer than we think.

For this it becomes necessary to use the ultimate simplification tool: example. So what are the practical examples of artificial intelligence? How often do we resort to its super powers in everyday life?

So many!

From moments of work to moments of relaxation, from finding solutions to the driving experience, from controlling the household robot that cleans our home while we are out, to online shopping.

Artificial intelligence: practical examples

As can be guessed from the premise, examples of artificial intelligence applied in everyday life are not uncommon. Let’s look at some of them in detail.

Entertainment and shopping

You know when you log on to Netflix to search for a new series and you are presented with a list of suggestions? These are the result of complex data processing that the platform derives from your behaviors, reactions as well as an analysis of your explicit likes. Based on this information Netflix creates content associations and then shows us “recommendations for you.”

E-commerce platforms that are able to guide our purchases by deriving information from our very behaviors are based on the same principle. You have surely happened to search for a product on Amazon and see not only purchase suggestions of items similar to the one you ordered, but also based on the choices of other users who used the same search keys.

In this sense, Amazon ‘s artificial intelligence is also able to predict, as well as obviously induce, our future purchases: did you buy a pair of hiking shoes? Why not also consider buying a backpack and water bottle?

What about voice assistants, which do all this and more? Alexa and Siri, to name the two most popular, are able to provide you with solutions on an increasing number of searches, as well as having an increasing number of integrations with services and online stores. In addition, they are a great hub for managing the home automation elements in your home.

On the road

Another tool based on artificial intelligence that is present in our daily lives is the navigator. It doesn’t just give us directions, but actual route directions depending on traffic, construction slowdowns, road closures, or vehicles stopped in the middle of the roadway. The navigator’s artificial intelligence can calculate the most suitable route for our needs.

And who has not heard of the almost limitless power of artificial intelligence employed in the features of Tesla, the all-electric car with learning capabilities that can make it autonomous in driving?

Communication and social

Going back a few years, the world of social and e-mail have introduced artificial intelligence capabilities into our lives. Who do you think filters the mail and deals with spam, or recognizes and places promotions in the appropriate folder? Who notifies you when an email goes unanswered for a few days. Artificial intelligence. Where do Facebook, Instagram, TikTok draw from to offer you sponsored content or ads close to your interests? From your own behaviors!

Artificial intelligence everywhere

And of course it doesn’t end there. For example, we can find translators using Artificial Intelligence; they are not always in line with the nuances of meaning, but still reliable and resolving in most cases.

Or again, going outside the perimeter of everyday life, it is possible to intercept the contribution of artificial intelligence in many other contexts, two among them being the military and the medical field.

The examples are many and across the board, and most importantly, likely to increase in a very short time.

What is the intelligent machine?

It is a computer that can handle a context of uncertainty according to the procedural patterns of the human brain:

  1. Gathering the stimuli of the situation he is in;
  2. By relating them to each other;
  3. relating them to any useful, previously stored data;
  4. Finding a solution.

Therefore, it is not based on programming but onlearning.

Applications of Artificial Intelligence in E-learning

A system capable of learning from experience is definitely an unprecedented resource for

E-learning

, albeit with the understanding that it is a dialogue in perpetual development.

The main current applications of Artificial Intelligence are:

  • Chatbot
  • Semantic Analysis
  • Big Data Management
  • Computer Vision

Chatbot

It is software that simulates a conversation. As you may have already experienced, it is mainly used by goods and services companies to manage online help or skim users who need assistance.

It acts as an intermediary between the user and the information, but it can also be used to collect data for later processing.

The chatbot uses Conversational Learning, which is a learning process that is developed through dialogue, becoming, in effect, a tutor who:

  • Is always available;
  • dialogues and teaches intuitively and cooperatively;
  • Provides information in a natural language;
  • He ponders the answer based on what he already knows about his interlocutor;
  • supplement data with external links;
  • Reduces fruition time;
  • offers statistics on questions, learning time, problems, interactions;
  • performs an engagement tracking.

Working on the capabilities of chatbots and combining them with machine learning skills, several types of text-based generative AI have been realized in recent years. Among them, the most famous case is surely that of Chat GPT , which began as a conversational chatbot but soon demonstrated its enormous potential in E-learning.

ChatGPT is not the only chatbot of this type, you can find many others including-to name a few-Google Bard, Sage or Jasper.ai

Semantic Analysis

Once information has been gathered, it is important to know how to handle it. Semantic Analysis is a way of processing written and oral texts, short or long, that mimics human cognitive processes.

Through software, key concepts can be extrapolated, quizzes can be generated, answers can be evaluated, and feedback can be provided where necessary.

In addition, the study of the semantic field enables:

  1. derive information from user input (content to be used for its own education, and to respond in increasingly natural and “human” ways;
  2. Provide machine translations.

Big Data

When the amount of data to be handled is so large that it becomes impossible to process it through human processes, we begin to talk about Big Data.

Their analysis is entrusted to algorithms that store, filter, search, transfer and analyze information, creating relationships and interconnections that improve the quality of the final response.

In the world of e-learning, leaning on Big Data analytics means accurately tracking the user experience so as to:

  • Monitor the user’s strengths and weaknesses;
  • Follow its path on the platform;
  • Isolate critical issues and take action;
  • Keep track of his or her schedule and preferred teaching styles.

Computer Vision

The acquisition of information by an Artificial Intelligence is not limited to text-type data but also to images and videos, through Computer Vision technology, which “teaches” computers how to read reality. Computer Vision represents the eyes of computers, and enables them to acquire data from digital images, videos, or through webcams, identifying the elements that make up the environment they describe, understanding their meaning, and reacting to their presence or alteration.

In all this, a key aspect is the speed with which these operations are carried out, which requires a great deal of computing power.

If this sounds like science fiction to you, consider that you have probably been dealing with it for longer than you think.
Have you ever noticed, for example, that your smartphone’s photo gallery has a search function that allows you to filter photos from one or more items? Here, that one is based on Computer Vision.

Moving to a higher level of application, however, you can think about how self-driving vehicles perceive the environment, the road, cars, people, and other types of obstacles in their path.

Among the many possibilities offered by Computer Vision we can find:

  • The decomposition of an imageinto different elements and the consequent possibility of classification of images in a database;
  • The detection of objects in an image or physical space acquired by webcam;
  • Real-time tracking of detected objects within an image or physical space.

These aspects can offer great possibilities, for example, if applied to a type of training that makes use of gamification, perhaps leaning on immersive technology such as Augmented Reality.

Artificial intelligence programs for E-learning

How can Artificial Intelligence come in support of E-learning? Soon said! In fact, there is a whole range of Artificial Intelligence software that can be useful for both those who implement training courses and those who use them. Let’s look together at some of the main software.

ChatGPT

We have already talked at length about ChatGPT, so we will just point out how it can be a vital aid throughout the production phase. Just using it to have a schedule to follow, to organize content, or to retrieve insights to go deeper makes ChatGPT a great ally for your E-learning. And consider that its potential is much greater.

ChatGPT has a free version, with limited functionality, and a subscription version that allows you to take advantage of all its skills

Jasper.ai

Similar in part to ChatGPT, Jasper.ai is an Artificial Intelligence program that performs copywriter functions, including SEO optimizations and copywriting skills for marketing. An ideal support not only to implement an E-learning path (at least in part), but also to accompany it to the market.

Grammarly

Once you have prepared texts for your course, you can get a grammar check through Grammarly, a free software program that, in addition to its generative writing features, allows for grammar reviews, text authenticity checks, and citation retrieval.

Midjourney

To manage the graphical component of your course, Midjourney is definitely an excellent resource. It is a ‘generative AI that creates images from a textual prompt, that is, a description of what you would like to use as an image. The more detailed the description and the more it moves away from possible ambiguities, the closer the final image will be to the desired one. Midjourney offers a trial, which includes a limited number of requests and binds the generated images to personal use only, and a subscription version that allows you to create as many images as you want and also use them for commercial purposes.

There are many alternatives to Midjourney, more or less effective, but having to choose one among them all, the choice would probably fall on Stable Diffusion.

Freepik AI Image Generator

Alongside the perhaps better-known Freepik, a database of photos and illustrations to enrich one’s courses (some free, some for a fee) there has also recently been Freepik AI Image Generator, an image generator addressed via prompts (as with Midjourney) and which again offers a free version (allowing a maximum of 3 prompts per day, each generating four images) and a premium version (increasing the number of daily prompts to 30).

Lumen5

Lumen5 is a free AI software that allows you to create videos, including making use of the many templates available.

Descript

Descript is text voiceover software, which you can train with your own voice (or the one you have chosen for your course) to quickly create any kind of audio stream to accompany your course (for slides, for dubbing video clips, or even for a podcast).
Descript also offers a free entry level and a number of pricing plans that affect the final quality of the audio, the level of editing operated by the software, and the potential of other features.

About the algorithms.

There are a number of toolkits that allow you to integrate your platform with machine learning libraries and software that allow you to build algorithms that analyze the data extrapolated during user behavior tracking in real time. This data can then be used to make the training more immersive and personalized.

These include TensorFlow (which is also open source) and ai-one.

What about accessibility?

Userway is software that operates on the cloud and leverages AI and automation to create accessibility solutions, improve user experience. It offers both a free version and subscription formulas with different pricing plans.

Artificial intelligence and E-learning: pros and cons

Like any partnership, the one between Artificial Intelligence and E-learning is also full of positive elements. From the very beginning, the following advantages emerged:

  • Being able to answer users’ questions in real time;
  • Interact with the virtual teacher in their native language;
  • Ensure greater accessibility for users with disabilities;
  • Create new custom content based on previous performance;
  • Use the collected data for accurate follow-up;
  • rely on greater objectivity and accuracy in exposure.

However, these pros also bring with them a number of cons:

  • Still high costs;
  • more impersonal interaction, limited and confined to education;
  • The risk of massive data loss in case of failure;
  • issues related to network accessibility and performance.

Over time, it has become evident how the use of AI can facilitate the process of designing and writing an E-learning course, quickly building a base from which to develop and localize a course.

As we have seen in recent times, then, Artificial Intelligence has also entered the content production sector in a big way, both in the textual and visual (not only images and graphics, but also video) and audio components.

This aspect is currently one of the most controversial ones, as it involves other people’s creativity and work, moving the issue into the hostile territory of copyright; two famous cases from this first half of 2023, for example, have been the rising of shields of illustrators and designers against graphic AIs such as Midjourney, and the strike called by Hollywood actors and screenwriters over the regulation of generative AIs in cinema.

Will such issues be found in the E-learning sector as well? For the time being it would seem not, but it may still be too early to tell.

Machine Learning and Deep Learning

So far we have only talked about Artificial Intelligence. Terms such as Machine Learning and Deep Learning are mistakenly used as synonyms.

The diagram below will help you gain clarity:

As you can observe, each constitutes – from time to time – the subset of the other.

What is Machine Learning

This is the general term for the process by which computers learn from data, starting withBig Data processing.

Algorithms used in Machine Learning can perform specific tasks without being programmed for that purpose. How do they do it?
Basically, by analyzing the data obtained they are able to identify patterns, derive predictions and improve their performance.

Of particular interest are artificial neural networks (ANNs), used by Deep Learning to replicate the behavior of human neurons.

In this area, too, a few examples may come in handy.

Machine Learning in Everyday Life

As we have seen, Machine Learning constitutes a branch of artificial intelligence, which is why some of the examples given above may also prove valid here.

A first common example certainly involves voice and image recognition. In the first case, Machine Learning causes spoken words to be transformed into text. The software, in this case, can convert the audio file to text files. Non-voice message lovers, thanks to this feature, can communicate in writing without pressing a single button other than the voice command button.

Also well known is the use of Machine Learning for image processing. Think of the use of facial recognition, for example, used as a way to access or unlock digital devices and software.

As mentioned earlier one of the most sophisticated uses of artificial intelligence is in the medical field. In this area, the contribution of Machine Learning is crucial in determining the diagnosis of certain diseases. Through the analysis and comparison of specific parameters, thanks to this technology, it is now possible to derive predictive data with respect to the progression of diseases or the possible outcomes of a specific therapy on a patient.

Even in the financial field, the application of Machine Learning proves to be no less. Consider, for example, the use of it by banking institutions with respect to customer behavior. Machine Learning enables banks to conduct market research and profile customers’ buying habits in order to predict their decisions. In practice, it facilitates them in taking the actions and initiatives of individuals.

What is Deep Learning

We can say that these algorithms are the result of theevolution of Machine Learning algorithms. Their use is leading to the creation of applications that were unthinkable just a few years ago.

In what fields are they used?

An example known to all is the autonomous driving experimented on Teslas. In this case it helps cars identify certain objects on the road, namely obstacles, pedestrians but also signs.

Deep Learning in Everyday Life

Even in the case of Deep Learning, which is basically nothing more than an advanced level of Machine Learning, it is possible to trace examples in previously mentioned areas.

One popular use of Deep Learning is to apply filters to photos. Many Apps make use of filters and features based on this technology to modify images in real time according to the user’s taste.

The music platform, Spotify, following the same logic as Netflix, uses Deep Learning for listening suggestions to users based on preferences expressed through song or podcast selection.

Perhaps less common, but very interesting, is the application of Deep Learning in software that converts handwriting to digital writing. Indeed, the need may arise to retrieve the contents of manually written documents and then save them in digital versions. Just think of the resources saved in terms of time and energy!

Machine Learning and Deep Learning in Digital Education

The technologies just described provide innovative and extremely useful applications in the field of digital education.

Here are some particularly interesting applications:

  • Content automation through quizzes to estimate one’s preparedness for a course, adjust the content offered, and improve engagement.
  • Collaborative learning, grouping participants according to skills, strengths and interests and avoiding conflicts.
  • Simplification of routine tasks so that teachers can devote more energy to the learning experience.
  • Personalization of content through semantic and Big Data analysis. It is most useful when applied to scenarios or gamification, where the interaction is adjusted to the user’s strengths and weaknesses.
  • Intelligent chatbot-mediated responses and feedback . Imagine that you have many people enrolled in a course and each of them has one or more questions to ask. How much time should you devote to this task? So much. Too much. To the point of risking overlooking some requests or having to respond hastily. A chatbot’s support might initially appear too limited and similar to an auto-responder, but it would gradually build on the feedback received, improving over time.
  • eProctoring to monitor exams online, taking advantage of webcams and microphones, inhibiting browsing to external sites and use of specific applications. Machine Learning helps confirm the identity of participants and recognize possible misbehavior.

At this point we see how these applications are used in two of the most popular e-learning platforms, Moodle and Docebo.

Moodle: from quizzes to learning analytics

These range from simple self-assessment quizzes to software that detects plagiarism in student work, to the more advanced Moodle Learning Analytics, which leverages Machine Learning algorithms to make learning predictions based on the activities performed by users. To do this they analyze:

  • The content viewed by the individual person (even multiple times);
  • Activities and discussions in which the user has participated.

These analyses, just stated, can be applied to the entire site or to specific courses and participants.

Among the many innovations developed on this LMS, one, currently in the works, is one for the integration of Moodle and ChatGPT with the aim of having an internal tutor who can help users who encounter difficulties while taking a course

Docebo: coaching and search optimization

Docebo has long offered several AI-based features. Here is a brief roundup.

  • “Virtual Coach”: harnesses the power of conversations with users to guide them through training activities and help them work within the platform;
  • “Advanced AI-powered search”- an internal search that analyzes any content on the platform. In the case of videos, for example, the AI directly considers spoken content;
  • Self-Tagging and Skill -Tagging: labeling resources, contributions, and skills to identify key phrases for their cataloging and searching;
  • “Invite to Watch”: a function applied to social learning that recommends users the ideal audience for certain content;
  • Tips for Admins and Managers: skim and reach out faster to users who might be interested in the course;
  • Suggestions for Students: guide them in viewing the potentially most relevant contributions and help them in creating a personalized training path.

Integration with Edugo.AI artificial intelligence software is recent, with the aim of providing the platform with the potential of generative AI.
This will allow the platform, among other things, a more personalized training experience in terms of content, simulations, and feedback, tailored to the user and his or her actual skills and knowledge.

Future Speculations: Artificial Intelligence and E-learning

Is the future of Artificial Intelligence applied to E-learning here?

Not yet. At present, AI is not able to replace the trainer, but it can make the trainer’s job easier by managing the more routine aspects of the job. Its potential is growing by the day, and new software is already proving to be a valuable aid to anyone working in education who has decided to give AI a chance.

However, at present, we do not think, that Artificial Intelligence can completely replace the trainer and provide a creative and engaging e-learning experience.

Part of this is also due to the regulatory issues of this technology, whose problems are growing in tandem with its benefits: copyright infringement and work ethics issues are only the most important problems that have emerged in recent times in relation to the use of AI for work purposes.

The main need remains that of regulation: once this fundamental juncture is clarified, we like to think that future developments of this technology might even accelerate, becoming a key tool along all production steps, but still as a technology in function of specialists and never as a replacement for them. Whatever the future developments, which we will be ready to document for you!

share on

Others like you have read:
blended learning
E-Learning
Blended Learning: how to make corporate training effective

Reading Time: 12 minutes That E-learning has changed the way training is a well-known fact by now. Since its establishment, however, two opposing schools of thought have stood out. The former has given full support to online learning, proving itself open to the novelties and countless possibilities offered by technology. The second, however, has distanced itself from it, continuing to prefer traditional learning modes.

That is, until attempts were made to make them coexist by making courses that could take advantage of the strengths of both. After numerous attempts, and with the support of increasingly developed technology, the potential of this methodological approach became apparent. Blended learning has thus begun to be talked about.

Read more
webinar tecnologie cognitive
Webinars
Webinar.The Real Impact of Cognitive Technologies in E-Learning.

Reading Time: 3 minutes Is Artificial Intelligence really changing the way E-learning is done or is it just an illusion?
If it is not, how can it be effectively integrated by making a real impact in training?
Together with isEazy we will address these questions during the upcoming webinar, with the aim of exploring the subject matter analytically and critically.

Read more
Webinars
Webinar. The ingredients for starry Digital Onboarding: the Cannavacciuolo case

Reading Time: 2 minutes We are ready for a new webinar by Frog Learning. How about you?
In the next appointment we will explore digital onboarding, a key process for the effective integration of new employees. We will tell you how an innovative and well-structured onboarding plan can generate positive effects in terms of retention and retention of new talent. To do so, we will look at the case history of Cannavacciuolo Group.

Read more
Sign up for our Newsletter

Take the Leap Forward in E-Learning training for Companies!

Logo frog learning