In this article written by five hands, 4 human hands and one AI, Learn Assembly invites you to a debate on the future of digital learning.
The cover image of this article was generated by AI.
The advent of artificial intelligence (AI) is transforming many industries, including digital learning. Digital learning players, who have long played a crucial role in the design and delivery of e-learning solutions, now find themselves at a crossroads. AI, with its ability to learn and adapt, presents both opportunities and challenges. Let’s take a look at five arguments in favor of digital learning players disappearing because of AI, and five reasons why this scenario might not materialize.
Antoine and Rémi: nothing to add to this introduction
1. Chat-GPT: Automating content creation: AI can generate bespoke instructional content at a speed and scale that human agencies can’t match, making the manual process of content creation obsolete.
Antoine: On paper, I agree, but experience shows that to create instructional content using AI, you need raw material, and what’s more, material specific to a job, a company or an industry: detailed PowerPoints, text, precise syllabuses, summaries of interviews with experts. However, such content is often either non-existent or fragmented. In such cases, how do you feed AI? AI will lead instructional designers to interact differently with experts. Some AI tools offer tones (academic, informal, professional): all texts will look the same, become smooth.
Rémi: Today’s AI tools can’t guarantee the veracity of the content provided (risk of hallucination). And writing content requires making choices (perspective, point of view, etc.). By automating content creation, aren’t we depriving ourselves of these choices, or limiting them? Finally, with AI, we can talk about the quick creation of a first version of content, which will have to be checked by humans (verification of data, reasoning, etc.). Talk of “content creation automation” seems premature. So I don’t completely agree.
2. Chat-GPT: Personalizing learning: Thanks to data analysis, AI can offer a highly personalized learning experience, adapting to learners’ individual needs better than standardized agency programs.
Antoine: Here again, on paper we agree, but really on paper: personalization comes up against IT barriers in large groups (security, foreign LLMs, GDPR) that slow down the adoption of AI. The development of customized engines will take time. The scaling-up of adaptive learning is not coming up against a technological barrier, but a human and IT one.
Rémi: For the learning experience to be personalized on the fly, it will be necessary to remove the famous risks of AI error. At this stage, this can only be done by humans. The promise of automated adaptive learning therefore comes up against the limits of current AI solutions.
3. Chat-GPT: Cost reduction: AI can significantly reduces the costs associated with designing, updating and distributing learning programs, calling into question the economic viability of the players involved.
Antoine: Customers are bound to challenge their service providers on price and want to internalize. However, on paper, internalization is an obvious option, but – as always – it comes up against reality. Today, the digital learning manager profession is in short supply. Talent is scarce, and these are demanding profiles who can remain freelance and live as digital nomads. The attractiveness of large groups has declined, and the remuneration packages offered by training organizations are too low (30k€ / year on average according to our barometer). This reality of the working world is going to slow down the desire to re-internationalize.
Rémi: We can anticipate a drop in costs and, at the same time, an increase in demand for adapted and rapidly operational digital learning, in a logic of immediate reaction to identified needs. In any case, this will drastically increase the expected ROI on learning needs that could not be covered by lengthy and costly e-learning developments. This could lead to an increase in demand for microlearning (at the risk of losing sight of the big picture).
AI will contribute to a significant improvement in the turnaround time for these systems. In the long term, we can anticipate an increase in the volume of projects combined with a reduction in costs per project. This will increase concentration in the digital learning market. It will become essential for these structures to reach a critical size in order to effectively manage the growing volume of projects and customers.
4. Chat-GPT: Real-time updates: AI enables continuous updating of instructional content, adapting rapidly to new information and technologies, a difficult task for agencies to achieve because of production deadlines.
Antoine: That’s not really true. Modifying a PowerPoint presentation, a video or any content on an authoring tool, with or without AI her AI or not, is quick and depends more on the tool’s ux than anything else. Only scormed modules remain hard to modify: everyone involved in digital learning has been pleading for ten years to get rid of scorm. Also, we run the risk with AI to have disposable modules, a bit like fast fashion.
Rémi: The technological building blocks are clearly lacking. As for letting an AI adapt and update learning content without any control, it seems risky to say the least, for the reasons already mentioned. The fact remains that continually updating training content can meet real needs in certain fast-moving technological fields.
5. Chat-GPT: Increased accessibility: AI can make learning more accessible, breaking down geographical and financial barriers, which could reduce the need for specialist services.
Antoine : If we’re talking about accessibility in the sense of RGAA (Reference Document for the Improvement of Accessibility) standards, then not at all. We’re back to magical thinking with this idea of “breaking down geographical barriers”, as in the days of MOOCs, which were going to revolutionize education and are now paying EdTech startups managed by private equity funds that are far removed from the humanist spirit of the early days. People who don’t have access to the Internet and who have other realities won’t have any more access to education than before.
Rémi: It’s one thing for AI to facilitate content creation by reducing costs, but I agree with Antoine on the reality of the digital divide: AI is not inclusive. What’s more, if we don’t transform practices enough with digital learning alone, there’s no reason to do it any better with mass-produced AI-based digital learning.
1. Chat-GPT: Human touch: Agencies offer irreplaceable human interaction, essential for motivating learners, cultivating critical thinking and encouraging collaboration.
Antoine: Learning consultants and instructional designers will become interviewers, journalists and workshop facilitators, helping to identify knowledge and the best learning methods, and managing production. Their role will remain essential; the job of instructional designer or digital learning manager in the sense of content production, on the other hand, will undergo major changes.
Rémi: Economic gains on the development of instructional resources can be redirected towards synchronous learning sessions, which are the only ones capable of fully contributing to the transformation we expect from successful a successful learning program.
2. Chat-GPT: Deep understanding of needs: Digital learning professionals bring a nuanced understanding of instructional objectives that AI can lack, especially in complex learning contexts.
Antoine: 100% agree. Understanding a subject isn’t just about creating content, it’s about living it, experiencing it, touching it. A profession, a notion, is lived in the context of people and in the context of work.
Rémi: The added value of service providers already lies in their ability to understand the hidden challenges facing organizations, to define relevant learning objectives on the basis of operational needs, and to remove obstacles to individual learning. Their position and image will be strengthened in the future.
3. Chat-GPT: Creativity and innovation: Digital learning players can innovate and create unique learning experiences, a capability that still eludes AI, which focuses on existing models and data.
Antoine: Until when? Tomorrow GPT will be able to create immersive modules, or connectors will exist between different technological platforms.
Rémi: Today, it’s clearly difficult to produce innovative AI content. But not all digital learning agencies offer unique learning experiences, because the pressure on prices and competition of recent years has gradually led them towards a form of optimization and industrialization (templatization, standardization of formats, etc.): the battle against AI will be hard to win in this case.
4. Chat-GPT: Ethics and privacy: Digital learning players can navigate ethical and privacy considerations in a more nuanced way than AI, protecting learners’ rights and data.
Antoine: The issue of data confidentiality will probably have a dampening effect on the use of AI until private AIs are developed. Companies are going to train their employees on this subject to avoid abuses. The question of sustainable design is also one of the subject’s blind spots: to date, AI is not in line with the challenges of frugality and digital sobriety.
Rémi: Many customers will be reluctant to accept the use of AI to process internal content and data. I think some customers will quickly prohibit the use of AI to generate content based on confidential data.
5. Chat-GPT: Support and coaching: They offer tailor-made support and assistance that AI cannot provide, particularly in solving specific problems and personalized coaching.
Antoine : I agree 100%, which is why Learn Assembly decided four years ago not to be a digital learning player, but a strategy consulting firm specialized in learning, precisely to anticipate changes in the market. And we don’t regret it.
Rémi : Right! The contribution of the digital learning market, which is currently very focused on instructional design and production, will gradually shift towards a more suited and individualized design (largely supported by AI tools) and towards assessment in the broadest sense, as well as a reinforcement of synchronous learning programs.
In conclusion, although AI represents a great way to innovate and be more efficient in the field of digital learning, the death of specialized agencies is not inevitable. The coexistence and complementarity of human and machine could well be the key to an enriched educational future accessible to all.
Antoine: Consensual conclusion by GPT-4, which seems to have political communication qualities in addition to its technological ones…
Learn Assembly is a hybrid consulting firm created in 2013 to support the transformation of all those involved in learning and employment. Our mission is to help them play a strategic role in their organizations to meet the challenges of skills in a context of environmental transition and technological transformation. We support the general management and L&D departments of major groups, public bodies and higher education institutions in their strategic development.
