The Rise of Artificial Intelligence in L&D: New Opportunities and Challenges

Our lives are increasingly powered by Artificial intelligence (AI), a silent companion that has been quietly integrated into our world since the early 1960s. Whether it's the facial recognition on your smartphone, email spam filters, or the weather forecast on your favorite app, AI and AI-reminiscent processes have seamlessly become an intrinsic part of our daily routines.

However, as AI extends its influence into various sectors, including the learning and development industry, it brings as many new ethical dilemmas as it does fresh and unprecedented opportunities. This article will explore some of these complexities as we explore ways to leverage AI in our learning programs.

AI: From Silent Agent to Principal Player

In many current applications, AI is merely a software agent, a piece of software that assists users or other applications while operating autonomously. It demonstrates a certain level of intelligence by simulating how a human brain would approach a problem or situation to generate conclusions. AI is typically placed in an environment and trained to observe, make decisions based on those observations, and execute actions based on those decisions to achieve desired outcomes.

For instance, Siri, the virtual assistant for Apple devices, can surprise you with an alert suggesting leaving home earlier for your upcoming meeting across town due to unusual traffic. You don't need to ask Siri to do this because, as an agent, it works behind the scenes. It observes your calendar, your location, and your destination. It decides to check the traffic and, depending on whether it affects your travel time, it acts by notifying you to leave earlier.

Nonetheless, AI is not always a silent agent. Nowadays, it has taken on a more active role. AI chatbots like ChatGPT and Bard can engage in human-level conversations and assist with complex tasks like coding and creative writing. On the other hand, image generation tools like Midjourney and Stable Diffusion can produce high-quality images and art-like pieces from textual inputs.

As you can see, AI can learn to act as if it were thinking like a human. But how does it know if it's getting it right? The answer is "machine learning," an algorithmic system that allows AI to improve itself through feedback. This feedback can come from a human operator, who may give a rating after, for example, a chatbot interaction. Feedback could come from self-assessment if the AI itself is trained to make such observations. Alternatively, feedback could be derived from a combination of both methods.

READ MORE ON AI | ‘3 Ways to Leverage an LMS to Promote the Proper Use of AI Tools Among Students’

AI’s Challenges for L&D: New Ethical Considerations

Policies and laws often lag far behind the introduction of new technologies. With the rise of artificial intelligence, and in the absence of widespread controls on its use, organizations will have to be aware of the ethical implications of working with the technology in two key areas.

1) The Importance of Striving for Unbiased Data

When we introduce AI into an environment, we provide it with a goal and a vast dataset to work from. It then processes the data through algorithms to determine the most effective way to achieve that goal. It's essential to recognize that algorithms aren’t as critical as the data for AI. Since data serves as the foundation of AI’s learning, it’s the data that leads to either positive or negative outcomes.

In 2016, an investigation conducted by ProPublica alleged that a computer algorithm used by judges was recommending harsher sentences for black individuals solely based on their race. Subsequent investigations suggest that algorithms can’t inherently exhibit discrimination. Instead, they’re influenced by the data they are trained on. Consequently, it was concluded that the AI used by the judges must have been trained on data that was inherently biased against minorities.

2) Intellectual Property (IP) and Data Collection

We've discussed how AI learns from data, but AI companies often don’t disclose the specific data used to train their models, giving us only a broad idea of what their dataset includes. ChatGPT was trained on a 300 billion-word dataset, while Midjourney used hundreds of millions of images. Both were sourced from publicly available online resources—an impressive amount of information, but the internet has plenty of copyrighted material. So, did these tools use that material indiscriminately?

OpenAI, the company behind ChatGPT, hasn’t yet provided a clear answer to this question. Meanwhile, the CEO of Midjourney acknowledged the use of copyrighted material. Regardless, both companies now face similar legal challenges, with numerous writers filing lawsuits against OpenAI and artists doing the same against Midjourney and Stable Diffusion.

And there's another concern: What happens with the information you share with the AI? Is it used to train the algorithm? ChatGPT acknowledges that it does use this data but provides users with the option to turn off the training of their model. However, in many cases, you may need to conduct thorough research to find this information, sometimes the answer is hidden within the software’s terms and conditions. The need for clearer transparency on this subject is a noteworthy issue.

It's important to remember that the safe and ethical use of platforms and tools ultimately depends on a user's intentions. The ethical considerations mentioned above, and others that may arise in the future, must always be a priority for responsible corporations. However, if a corporation is genuinely committed to fostering innovation, it should also strive to harness the potential of this technology for positive purposes.

3 Ways to Leverage AI for Digital Learning

When used responsibly, AI can be an incredibly powerful tool. Instead of hesitating, companies should proactively develop policies for safe engagement with AI. Otherwise, they may allow competitors who understand how to harness AI to gain a significant competitive advantage by forfeiting this opportunity.

Let's explore the various ways AI can enhance Learning and Development (L&D) programs:

1) Accelerating Course Development

AI expedites tasks that typically consume a significant amount of time by providing us with groundwork to build upon. Thanks to that, it can streamline the entire course or training design process as follows:

• Chatbots can assist in structuring the flow of your new program.
• Chatbots can assist with writing initial drafts of course text and video scripts.
• Text-to-speech tools can be used to create audio narrations for your videos.
• Text-to-image tools can generate avatars and graphics for multimedia elements.
• Chatbots can enhance the quality of your materials by offering feedback and suggestions, potentially rewriting sections for improved readability, removing redundancies, or even detecting incorrect information.
• Chatbots can also generate questions and responses for assessments and provide feedback for learners.

It’s important to highlight that AI content generation relies on your feedback and guidance to produce results aligned with your objectives. Otherwise, it may produce generic outputs. To maximize its performance, invest time in crafting specific prompts, ideally consisting of two to three paragraphs. Providing context, such as a role to portray, a clear objective, and the target audience, leads to superior results.

2) Enhanced Accessibility

Accessibility is a legal requirement in higher education and the workplace. However, historically, achieving accessibility has been a time-consuming process. That may be why institutions often overlook the requirement (an issue underscored in a 2023 letter from the U.S. Justice Department and the Department of Education).

Fortunately, AI has the potential to help us tackle this issue due to its ability to operate faster than humans. For example, students with hearing impairments typically require subtitles or transcriptions for video content. A trained transcriptionist may spend four times the duration of a video to transcribe it accurately. If a course heavily relies on video content, creating transcriptions would require a significant amount of time and compensation for the transcriber.

Now, AI with speech recognition can perform an initial pass of the same job, providing timestamps and speaker differentiation in less time than the video's duration—a significant improvement. This not only helps the transcriber but also enables non-trained individuals to create the transcriptions themselves.

There are also AI tools for PCs and smartphones that can provide spoken descriptions of images displayed in other applications. These tools could be beneficial for students with vision impairments, enabling them to better understand images within your course content on both the mobile and desktop versions of your learning platform (though accurate image alt text should additionally be created as part of your process).

Furthermore, tools exist that allow you to build an AI version of your voice by simply uploading audio recordings of yourself. This AI-generated voice can then be used to provide audio versions of your learning content for visually impaired individuals, eliminating the need for studio recording and extensive editing. While it’s true many students with the kinds of accessibility needs already use text-to-speech software that allows for computer-generated narration, having the teacher's voice—a familiar voice—has the potential to be more engaging for the student.

3) Roleplay Simulation

One potentially innovative role for AI would be to set up role-play scenarios where chatbots take on roles in a specific interaction, simulating knowledge and behavior to facilitate learning or skills training.

For instance, if you have employees who frequently travel, AI could act as a travel guide, providing information on cultural insights, customs, and local norms to prepare your staff for their destination. Similarly, AI could serve as a tutor for employees learning a new language, offering feedback in the target language.

Moreover, an AI could act as a potential client for sales training by providing prompts that define the client's profile, including personality traits (e.g. demanding or indecisive) and their position within the sales funnel (whether they are considering the brand or ready to make a purchase). It's important to emphasize that the use of data is crucial for this type of roleplay. The more information the AI has about your customer's profile, the more accurate its personification will be.

YOU MIGHT ALSO LIKE | ‘Creating Real-Life Learning Experiences: Augmented Reality in Education’

What’s Next?

AI brings forth a wealth of exciting possibilities and, undoubtedly, some significant challenges. As discussed in this article, we encourage you to embrace the boundless potential of AI while remaining mindful of its ethical dilemmas. This way, you can find responsible ways to leverage its transformative capabilities.

If you'd like to delve deeper into this topic, we invite you to explore our academy and enroll in our free course, "The Rise of AI and ChatGPT." This course covers all the features mentioned in this article and provides new stories and use-case scenarios to comprehensively understand the subject.