The Development Of Our Writing
(We = non-native English bilinguals)
How does the history of our English text editing go?
First, there was a thick five-kilo dictionary and non-billable proofreading hours from our linguist friend. Then, there was an online search engine with copy-paste and search functions. Then came online dictionaries and thesauri, minimizing the need to get up from our chairs as we’re writing. At some point in time, text editors also introduced spell checkers that actually made sensible suggestions, but some other, external ones also appeared. Some kinds of “smart” apps.
And then…generative AI took over the world.
All the editing and proofreading roles were announced to become extinct.
We became free and independent top-tier content editors.
And alright, all is well, as this is just cosmetics. However, what happens when we talk about the history of content creation in moments when you have no inspiration, time, or confidence?
It could go like this: as an initial boost, maybe we had a person to talk to or a teacher who could nudge us in the right direction. For a “more thorough contribution,” there may have been a class geek who was kind enough to write a paper for us. Or perhaps our older sibling or (less likely in those times) an aspiring parent. Research—for some, too tiring; for others, the well of knowledge: libraries. And then, the internet opened its door to innumerable possibilities for finding inspiration. We still had to dig deep, though. It was time-consuming.
And then…generative AI took over the world.
Suddenly, there is the possibility to reach productivity unparalleled to any of the earlier historical stages of our era.
With this tool and maximized productivity, could we be called top-tier content creators?
The Risk Of Using AI In Content Creation
You’ve probably been using AI long before the hype was raised around generative AI. Most likely with audio/video streaming services, social networks, or online buying platforms. I have, at least. And the question of whether and how we should use AI is, as with all things new, the typical Plato-cave question. Either you’ll stay in the cave, blind to any catharsis (and also danger), or you’ll venture out of the cave to experience enlightenment.
More or less, the general feeling is that we should use AI for good, for improvement, and educate ourselves on how to use it to minimize potentially negative effects. Sensible, although we have to be aware of the difference between generative AI and regular, weak AI.
Weak AI seems safe, but reaching for generative AI is tempting and bold. I completely agree that it can be helpful and time-saving. In this context, it’s an upgraded search engine, text editor, and assistant, and undoubtedly many positive outcomes can be detected, even for businesses.
At this point, however, this is where I would draw the line. In terms of content creation with GenAI, the questions we, as eLearning content creators who make a living out of it, should be asking are the following:
- The greatest concern is that if I let GenAI create for me, how genuine is that content?
- Does it respect the work and authorship of others?
- How do we know the displayed content is legitimate?
- Ultimately, what is the source of information displayed on our screens?
I’ll try to answer these in the sections to come.
Generative AI, Plagiarism, And eLearning
Generative AI is trained on existing materials, and I assume there is only a certain amount of its capacity to produce content. The main doubt is whether this content is genuine. These models are said to be pre-trained, so what is in their raw training materials should or could be reflected in their answers. So, we’re witnessing many biases in some interpretations by GenAI.
We can’t even say how respectful of others’ work GenAI is. The thing is, in the academic and professional community, I’ve been taught to always pay respect to others’ work, even if it has been used for inspiration, as a basis for my own writing. Anti-plagiarism policies and practices, aligned with copyright laws, are nurtured globally, and their importance has been raised to the regulatory level. In other words, they’re not here just for cosmetics.
Anti-Plagiarism And Copyright: What Does The World Say?
On a global level, the world is united around the idea that plagiarism should not be tolerated and that the protection of copyright is “crucial to intellectual creation,” the almost poetic formulation used in the EU’s InfoSoc Directive, Section 9.
The strength of each region’s regulation could be examined in more detail; however, even the blogging community strongly agrees that credit should be given to authors whose work has in one way or another contributed to the creation of ours. There is no official research to support this, but—hey—here is the idea for one!
Given all this, two perspectives may be given:
- Using GenAI’s text in our work should be communicated transparently, as I see no reason why we should not mention it as a source. Okay, I see one: if we used it for an entire piece—this is tricky then. On the other hand, are we plagiarists if we don’t mention it even in these cases?
- Using GenAI’s text in our work should not be done, because we are unfamiliar with its sources in the first place. It may seem radical, but it is what it is.
Back To eLearning And GenAI
We haven’t forgotten that all this is written from the perspective of an eLearning creator at an eLearning agency. As you may have concluded, the principle of avoiding AI-generated content would be applied here, too. At least, the working postulate could be the following: if we’re selling our authorship for a living, it would be unethical to smuggle GenAI’s work as our own.
GenAI working as a dictionary or language editor is fine, and so are similar tools that have made this their specialty.
But—and we suspect that this may be one of the next questions—what if we use it to set up some general processes, concepts, learning plan suggestions, or any other similar framework that is then filled with “meat” by ourselves?
I fear the day when clients get identical solutions from each and every agency. Whenever I think about using GenAI for the mentioned purpose, this thought spreads apocalyptically through my mind, draining any desire to use it.
I’d like to add a note to all this. Actually, it is an insight provided by Julie Dirksen in her blog post, “The Pesky Challenge of Evaluating AI Outputs”. I find it really interesting and worth the contemplation: if or when one gets output from a GenAI, how competent are they to judge its validity? Even on a weak-AI level, often suggested grammatical corrections are invalid due to the wider context and the machine’s inability to read and understand it. As a consequence, the tool is only useful if you yourself are able to scrutinize it.
AI Tools For eLearning Development
AI is not as black as painted, though. The eLearning production teams still enjoy some of the benefits of AI tools, as do our clients, so here are some examples:
- Text editors – Always good to have, and we wish we had better ones in Croatian, too. Would a client trust you if your writing was a grammatical calamity? We wouldn’t.
- AI voice generative tools – With time, they’ve become very realistic and represent a budget-friendly option for less demanding scripts.
- AI video avatar generators – In time, we trust they, too, will become realistic.
- Dubbing and translation – Immense time saver. Nevertheless, a high level of scrutiny should be provided in translations, as often context, sarcasm, and culture-related jokes can be missed in translation.
We admit we could be more open to GenAI, but as long as authorship is our livelihood, the most ethical approach is to steer clear of dubious sources. Efforts should be made to avoid plagiarism and fully respect copyrights. This principle is globally nurtured, and authors should get clear credit for their work.
In the realm of eLearning content creation, storytelling and creativity are still highly esteemed and encouraged, and we are still to be reassured if pure innovation can come from anything other than human minds.
What Does Chat GPT Have To Say?
Just for fun, to end the article in style, we asked ChatGPT, based on GPT-4 architecture, the following question: “In no more than 50 words, could you please mention what the potential risks of using generative AI for content creation in eLearning are?”
This is the reply: “Potential risks include quality control issues, ensuring AI-generated content is accurate and aligned with learning objectives, ethical concerns about data privacy and bias, and the potential loss of human creativity and empathy in educational content.” (This question and answer were provided on July 19, 2024.)
A bit later in time, Claude.ai gave us this output (February 4, 2025): “Key risks include potential AI hallucinations creating inaccurate learning content, copyright/plagiarism concerns, loss of authentic human voice, inconsistent quality, and difficulty maintaining brand tone. AI may also generate biased content or miss crucial pedagogical nuances needed for effective learning.”
eWyse
eWyse is an award-winning agency, using a unique methodology called the 3C Framework to help build perfect eLearning courses that engage, entertain, and educate learners while helping companies to reach their objectives. Let’s discuss your ideas!
Source link
