How to use this guide
- Yes, it’s a long read, but everything you need is right here. Think of it as a guide. You can come back to it whenever you like. Bookmark it.
- You’ll mostly learn about what AI can do for employee training. You will only get brief explanations on the ‘how’ (we’ll explore some use cases more deeply in other articles, let us know what interests you).
- The use cases for AI for learning are structured by your high-level goal (e.g. generate something new) and what Generative AI is really good at.
- To find your ideal use case, apply this framework to the major steps of the training design and delivery process.
We won’t get into the details of ChatGPT’s history or the technical side of AI. There are many places to learn about those things. For a quick overview, check out one of our webinars. But if you’re reading this, it’s probably because you want to use ChatGPT and AI for learning.
So let’s jump right in.
Framework: The 4 Key Strengths of Generative AI
We’ve grouped some ways to use AI for learning. These are based on what Generative AI does best and why you’d want to use it. Remember, these categories aren’t always clear-cut; they can blend depending on how you see them or define them. The categories come from lots of time spent using ChatGPT and other AI tools in our work. We’ve also learned from research and talking with our customers and partners.
1. Generate: When You Start From Scratch to Make Something New
As the term generative AI indicates, the main promise of this technology is to generate output. In this case, you typically start from zero on the green field. The page is blank and you want to create text, image, code, etc. that doesn’t exist yet. Example: you tell ChatGPT to write an email to your boss.
2. Repurpose: When You Already Have Material and Want to Make Something Different Out of It
Quite often, you do not want to reinvent the wheel. You may already have some material available that you want to work with. AI can take what you feed it with and transform it into a different format, style, medium – basically whatever you want it to be. Example: you improve the writing of your email or translate it.
3. Interact With Knowledge: When You Want to Learn More About a Topic
One of the reasons for the breakthrough of ChatGPT is that you can chat with the AI that underlies it. Chatting is one of the easiest ways to interact with a computer (the technical term is a conversational interface). Access to knowledge has never been more frictionless. Example: you research the French Revolution.
4. Sparring: When You Want to Improve with Feedback
ChatGPT and other Large Language Models (LLMs) have been trained on a vast amount of text data, incorporating many perspectives, including biases. Unlock these aggregated views by using AI as your sparring partner to gain a fresh new look at your work. Example: review your learning design.
Now apply this framework to the task at hand within the learning design and delivery process.
Major Steps of the Training Design and Delivery Process
Planning & Preparation:
- Lesson Planning
- Lesson materials
Interaction & Implementation:
- Monitoring progress
- Feedback students & pupils
Stakeholder Communication & Collaboration:
- Parents
- Internal (organization)
Evaluation, Reflection and Further Development:
- Follow-up
- Reflection
- Evaluation
When you match what the technology is good at, your high-level motivation of using it, and the task at hand, this will guide you to the right prompt or use case.
Where to start? ChatGPT vs. Specialized Tools and Integrations
You might wonder, where do I get started when I want to use ChatGPT and other Generative AI tools for learning and development? It comes down to two approaches, that you can pursue separately or in combination.
Use ChatGPT and other AI-Chatbots Directly
The first one is to use one of the most popular chatbots like ChatGPT, Google Bard, or Bing Chat.
They allow you to chat with the AI that’s underneath it. When we look under the hood, the foundational technology is called a Large Language Model (in short: LLM). Every provider of these chatbots builds on a different LLM:
- GPT 3.5 (free) or GPT 4 (paid) when you use ChatGPT by OpenAI
- LaMDA when you use Bard by Google.
- GPT 4 when you use Bing Chat by Microsoft
While ChatGPT currently gains the most attention, there are also other Large Language Models and Chatbots that are worthwhile exploring like perplexity.ai or Pi.
Expect the number of models you can choose from to increase over time. This includes open-source models, from which experts expect a lot. Meta for example has made their LLM called Llama 2 available as Open Source and free for private and commercial use.
Before picking the right chatbot and LLM for your learning use case, you’d actually need to evaluate which one allows you to tackle your challenge the most effectively and efficiently. As this is quite time-consuming and to keep things simple and practicable, we’ll focus on ChatGPT for employee training in this article.
Let’s say you are just getting started and don’t want to spend money right away, the free AI chatbot solution by OpenAI building on GPT 3.5 is a good start. As you’ll gain experience you will reach some limits of the model (remember, the LLM) and might want to think about an upgrade to the paid version, ChatGPT Plus, or a specialized tool (like eggheads).
To sum it up: start with your own free account on OpenAI here.
Use Specialized Tools and Integrations Into Existing Tools
The second way to harness the power of AI for training is through specialized tools that have integrated AI capabilities. Experts compare AI with electricity. This means: it will become part of every product and we won’t even notice or mention that the specific feature we are using builds on AI.
Frankly, this has already been the case for example when translating text or correcting the spelling in your word processor.
The current development goes in the direction the experts predicted. In fact, major software developers like Microsoft, Google, Adobe, Salesforce, and many others were among the first to move and launch new capabilities in their existing products. Microsoft calls their solutions copilots that you will find in Word, Excel, etc. Google highlighted certain features that help write an article in Google Docs or an Email in Gmail.
In addition, literally thousands of new or updated tools have emerged not particularly with a focus on learning but across the board. While the space is flourishing, time will tell, which of these solutions will constantly deliver value to their users.
In the learning space, major edtech companies like Duolingo or Khan Academy enhanced their solutions with AI (read more here).
To sum it up: there’s a tool for (almost) everything, focus on the goal you are trying to achieve.
1. Generate – make something new
This is the most common and straightforward starting point. You have an idea and type your instructions into the Generative AI tool of your choice.
Text
ChatGPT is built on what’s known as a Large Language Model (LLM). In simple terms, an LLM like GPT 3.5 or GPT 4 is incredibly good at generating text. The power of text-based AI is invaluable in many professional settings, including learning and education.
However, its wide range of applications can also feel overwhelming at times as these models can generate text for virtually any situation where words are needed.
Let’s look at some types of text you can generate with AI: Emails, letters, Poems, raps, shanties, dialogues, outlines, concepts, reviews, applications, online content for websites, blogs and social media, synopses, research proposals…
Examples in Learning
Stages 1 and 2: Lesson plans, class activities, icebreakers, discussion starters, exercises, work assignments, tasks, group work exercises, sample solutions, module descriptions.
Stages 3: reflection and further development: Email to subject matter experts, progress reports, structure parent-teacher conferences, update student files, project applications, evaluation reports.
When choosing an AI tool, you have options like GPT-3.5 or GPT-4 from OpenAI. GPT-3.5 is free, but GPT-4 comes with a paid subscription service, ChatGPT Plus. Other alternatives include Google Bard, Bing Chat, Perplexity, and Claude.
Getting the best results requires good prompts. We have a separate blog post dedicated to this topic. Additionally, some premium tools come with specialized features for generating specific kinds of text.
Recommendation
Treat the AI-generated output as a first draft. Always double-check the facts and be aware of any potential biases in the text. Editing and revising are crucial steps in ensuring quality and reliability.
Pictures and graphics
AI has not only become good at writing but also at painting and illustrating. Just as text-based AI has changed how we write, these picture-generating AI tools are transforming how we visualize information.
To achieve the most effective visual outputs, crafting the right prompt is essential. Writing prompts for AI image generators like Midjourney, Stable Diffusion or DALL-E can be a different challenge compared to those for text generators. For guidance, you can consult this resource or this resource to master the art of prompt design for visual AI.
AI image generators have gone beyond just creating new pictures. Together with design software like Adobe Firefly or Canva, they now offer a wide range of AI-enhanced editing features. These tools can also improve existing images by replacing or removing certain objects with the stroke of a brush (e.g. Inpainting by Midjourney or Canva) or expanding the background (Adobe Firefly).
Another approach is to work with ChatGPT Plugins. For example, there are some that research data and create graphics.
Examples in Learning
Basically, across the board. You can integrate these AI-generated images into various formats, such as presentations and letters, to enrich your content and engage your audience more effectively.
Recommendation
Currently, these AI tools are good at illustrating information or expressing moods. If you want to create an infographic or a statistical diagram, traditional means or specialized tools might still do the trick.
Words of caution: Image generators have become so good so it’s really difficult to find out if an image is real or fake. If you’re using these tools to create images, please do so responsibly to avoid misleading others. Additionally, be aware that there are ongoing legal debates concerning potential copyright infringements, as these AI tools have been trained on potentially copyrighted material
Computer Code
According to a McKinsey study, software development is one field where Generative AI promises to deliver significant business advantages. ChatGPT or specialized AI coding tools like Github Copilot, Codey (Google) or Code Llama (Open Source, byMeta) streamline software development by automating monotonous coding tasks, producing boilerplate code, and even assisting with debugging. Additionally, they have the capability to output code and text in structured formats, such as JSON, making it easier to integrate into existing software products. In this light, there is a capability of ChatGPT called “function calling”, which allows the generated output o trigger call functions in other software.
Examples in Learning
Most likely when you’re a programming instructor for different exercises or projects.
Other outputs
AI has matured significantly in the generation of text, graphics, and computer code—applications with broad appeal that can be used by almost everyone. In addition to these widely applicable use cases, AI also offers specialized solutions for niche use cases, although these may not be as mature as the more universally applicable tools.
Video
Video content is also within AI’s reach, with platforms like D-ID.com, DeepBrain, and Synthesia offering video generation capabilities.
Voices
When it comes to audio, platforms like Meta AI Voicebox, murf.ai, Speechify, and Play.ht offer synthesized voice solutions that sound increasingly natural.
Music
When it comes to music composition, Beatoven.ai and Google MusicLM are pioneering the way.
Excel
ChatGPT can generate Excel formulas and tables to assist with data management tasks.
Data sets
Additionally, ChatGPT can produce datasets that help in training other machine learning models.
3D
Even in the 3D world, tools like Nvidia’s Neuralangelo are making strides, allowing for the generation of intricate 3D models.
2. Repurpose – make something different
As a learning professional, you frequently get source material from subject matter experts or may already have existing resources to build upon. When starting from scratch isn’t necessary, Generative AI becomes an exceptionally powerful tool for repurposing your existing content.
Here are the key dimensions of your content that you can alter when repurposing: language, structure, and form.
2.1 Language
Large Language Models (LLM) like GPT 3.5 or GPT 4 know the statistical probability of how words correlate. Or in other words: they are masters of language. However, consider that these models are primarily trained on English data, making English their strongest language. The quality of their output diminishes for less commonly spoken languages. Developing LLMs for languages other than English remains a topic of ongoing research and presents a considerable challenge (Source).
With ChatGPT and similar tools, you can repurpose existing content on the language level in various ways. Specifically:
Tonality and Style
Let AI modify the tonality and style of existing content when you need to target a new audience or wish to cast the material in a different light. If you’re writing in a language in which you’re not fully fluent, AI can assist in enhancing the style. In fact, it can also be just for fun. Add instructions to your prompt about how you want the repurposed content to sound, turn it e.g., from serious to funny, from factual to flowery, or even into pirate-speak.
Complexity
Although complexity isn’t only a linguistic aspect but also a matter of the content itself, AI can be invaluable in simplifying the material. Content provided by subject matter experts may be too difficult and thus irrelevant to a general audience. AI can assist in breaking down complex topics into more digestible information. To do this, include specific instructions in your prompt such as “simplify,” “explain for a five-year-old,” or “explain to a friend at a bar” (you can also specify “after three beers”). Specialized AI tools for simplification are also available.
Translation
To connect with participants in different geographical regions, consider using AI to translate your existing content. Given the language constraints previously discussed, it’s advisable to treat the AI-generated translation as an initial draft. ChatGPT can be used directly for this purpose. Several tools, including our own microlearning solution eggheads, come with built-in translation features. This allows you to instantly translate your chat-based learning modules into multiple languages.
Grammar
Use AI to fix typos and grammatical errors. Even though spell-checking has been a part of word processing software for a long time, AI tools like ChatGPT can also do the job for you. In some cases, fixing grammar and spelling might just be an extra bonus you get from using your original prompt.
Politeness
In some languages like German, French, Spanish, Hindi or Japanese you address people you are familiar with differently than strangers. There is an informal and a formal way to address readers. If you need to adapt your text to a different audience, tell AI to do it for you.
Personalization
To engage your audience more effectively, consider speaking to them directly by using “you” in your text. AI tools can help convert text from a third-person perspective, like “employees need to,” to a more personal and direct tone, such as “you need to.”
Direct/Indirect Speech
For a more dynamic and engaging narrative, consider alternating between direct quotes and paraphrased content. AI tools can help you seamlessly switch from direct speech, like quoting someone verbatim, to indirect speech, where you summarize what was said. This variety in presentation can make your content more interesting and relatable to the audience.
2.2 Structure
Summarize
Yes, there’s just too much to read, listen to and watch out there. No wonder summarizing content is one of the most mentioned ways consumers plan to use Generative AI for.
To make lengthy content more accessible and digestible, you can use AI tools to create concise summaries. There are different approaches to summarize articles, blog posts, Youtube videos and podcasts.
To summarize plain text, paste it into ChatGPT. Instruct it to summarize the input or just add a “tl;dr” (too long; didn’t read) at the end of the pasted content. In addition, there are specialized tools like tldrthis.com, anysummary.app, and upword.ai or Chrome Add-Ons like Wiseone and Summereyes for summarizing web content. Scisummary or Summarizepaper summarize scientific articles.
Google experiments with a feature called “SGE while browsing” in its Chrome browser. When you visit certain websites, this feature displays a list of the main points of an article (source). You can also summarize and chat with Youtube videos directly with Bard (its AI-chatbot).
If you don’t want to watch an entire Youtube video, youtubesummarized is one possible shortcut or the Chrome Add Ons Eightify or “YouTube Summary with ChatGPT & Claude“. Snipd.com focuses on extracting the key messages out of podcasts. This allows you to deliver the essential information in a condensed format, saving your audience valuable time.
Data Management
Use AI to streamline and organize your data effectively. For instance, AI can sort your data by specific categories, group related items together, or even extract essential information such as “all questions” from a dataset. Specialized AI tools can also clean your data, removing any errors or inconsistencies, and sort it into a more useful or manageable format. This can be particularly valuable in tasks such as market research, customer segmentation, or any situation where refined organized data is crucial.
2.3 Form
When using AI to modify the form of content, we focus on instances where the initial input is already a substantial content resource, rather than a simple text prompt. This does not include text-to-image transformations, like those offered by Midjourney, Stablediffusion, or DALL-E.
Transcribe
When you use AI to transcribe, it converts spoken words into written text (speech-to-text). This is particularly useful for dictating notes into your smartphone or transcribing webinars and meetings. Various platforms like Zoom, Microsoft Teams, or Otter.ai offer this capability.
Image Recognition
Leverage AI for tasks like image-to-text conversion, which can encompass image recognition, QR code scanning, and Optical Character Recognition (OCR). This technology allows you to extract text or data from visual elements, facilitating a more seamless integration of image-based content into your workflow. In the educational setting, specialized apps exist that can scan and transcribe handwritten text from students, converting it into a machine-readable format for easier grading and review. By now, ChatGPT can also analyze images.
Voices (Text to Speech)
With AI-supported platforms like murf.ai or naturalreaders.com you turn written text into spoken audio. These tools can transform articles, books, or even your own written content into a listenable format, providing an alternative medium for content consumption. In a learning context, this can be applied to create audio-only materials like podcasts or to add voice-over to instructional videos. Recast turns texts into Podcasts.
Text to Video
Take advantage of AI solutions like Vidon.ai to transform text-based content into dynamic video formats. These platforms provide a new avenue for engaging your audience, turning written information into visually appealing and easily digestible video clips.
Video to Text
Use AI tools such as Veed.io to transcribe audio and visual elements from videos into written text. This enables you to create transcriptions, subtitles, or even full articles based on video content, expanding the ways in which you can repurpose and share your material.
Long Video to Short Form
Use AI services like Vidyo or Opus to condense longer videos into brief, focused clips that capture the essence of the content. This form of video summarization helps you present key points quickly and effectively, making the content more accessible for busy or less committed viewers.
3. Interact with Knowledge
The limits of Artificial Intelligence – and its “knowledge”
The term “Artificial Intelligence” (AI) often leads us to believe that the systems we interact with have some form of intelligence. While a deep dive into the concept of intelligence is beyond the scope of this article, it’s important to note that AI-based conversational agents like ChatGPT, Bard, or Perplexity appear intelligent largely due to their advanced language processing capabilities. As humans, we tend to mistake eloquence as competence.
Sure, it’s more than remarkable to see the wide range of tasks that AI can perform today. For example, ChatGPT has been reported to pass various exams (source). Interestingly, some people have applied the concept of IQ, which is designed to measure human intelligence, to ChatGPT. They estimate its current IQ to be around 155, close to Einstein’s 160. Given the speed of progress, these systems will only become better.
Still, this brings us to an essential point: AI’s “intelligence” is fundamentally different from human intelligence. While AI is becoming better at reasoning and making connections between different pieces of information, this shouldn’t be confused with human-like consciousness, self-awareness, or genuine understanding (at least not today). The reason is that AI algorithms, including Language Learning Models (LLMs) like GPT, generate responses based on statistical patterns in the data they’ve been trained on.
This limitation has significant implications, especially when using AI as a learning or tutoring tool. For instance, these systems can inadvertently perpetuate biases present in their training data or even spread incorrect information. Therefore, while AI can be a powerful tool for accessing and interacting with information, it’s important to approach it as a supplement to human expertise rather than a flawless knowledge database.
Or as ChatGPT says itself: I’m a large language model trained to assist with a wide range of tasks, but it’s important to note that I’m not a specialized knowledge database. My responses are generated based on patterns in the data I’ve been trained on, rather than accessing a database of verified information.
ChatGPT
Or Bard: I am not a knowledge database. A knowledge database is a structured collection of information, typically organized in a way that makes it easy to retrieve and use. I do not have a structured knowledge base, but I can access and process information from the real world through Google Search and keep my response consistent with search results.
Bard
Despite these limitations, AI still fundamentally changes how we access and interact with knowledge. Or in other words: search and learn.
The Dream of the Personal Tutor
Even though responses from AI-Chatbots like ChatGPT might be biased or incorrect, GPT-4, when adequately trained, can already serve as an excellent tutor. It shows remarkable potential for adaptive, valuable teaching. Visionaries like Bill Gates even stated during a keynote speech:
“The AI’s will get to that ability, to be as good a tutor as any human ever could”
Bill Gates
With eggheads, you can create a conversational tutor-like experience based on your own information.
A survey in Higher Educated, which involved more than 1,000 students across all academic levels, revealed a significant shift in the use of tutoring resource: 75 percent of respondents reported using ChatGPT as a preferred educational tool, while only 25 percent relied solely on human tutors.
Education is not going to be the same anymore. You will have infinitely patient teachers that explain the exact example that you want to know at every level, in every step, in every language.
Ludwig Ensthaler
Tools from Khan Academy (and some of our own experiments) suggest that GPT-4, when properly prepared, is already an excellent tutor.
So, what does it actually know?
As Large Language Models are trained on vast amounts of data, the output they generate is determined by the training material. In simpler terms: if the training data covered a certain topic in high and correct detail, the answers by ChatGPT have a very high likelihood to be correct.
For example, if you ask what the capital city of England is, it will say London, because this information has a high frequency in the training data. For us, it appears the system is intelligent, but it “just” provides that answer based on the statistical probability. For us, it seems that ChatGPT knows what the capital city is. As a rule of thumb, the more public and solid public sources were, the more reliable the response is. The more niche or private a topic is, the more likely is an incorrect or even hallucinated response.
Search and research
The rise of ChatGPT has triggered a “code red” for the world’s leading search engine, indicating a potential disruption in the search for information (source). While chatbots like ChatGPT are effective in answering some queries, Google & Co. still hold the advantage of real-time access to the internet. ChatGPT was disconnected in September 2021, so it does not include any information or updates post that date.
However, it remains a reliable source for general topics, offering an uncomplicated method to explore various aspects of a subject. Its ability to provide quick answers makes it a useful learning resource. However, for academic research, ChatGPT has been found to fabricate sources, hence specialized academic research databases like Silatus, Semanticscholar or Jenni remain a more credible choice.
Chat with Knowledge
ChatGPT has transformed knowledge from being static to interactive. Instead of reading a page in a book or a website online, you can now access information interactively. It allows follow-up questions, and provides further explanations and comparisons. This offers a great opportunity for self-motivated learners looking to explore specific topics.
However, for those requiring more guidance, it could be overwhelming as they might not know what to ask. In such instances, asking ChatGPT for follow-up questions can be helpful. If you are looking to chat with knowledge that’s not publicly available or was published after September 2021, there are specialized tools available that allow interactive conversations with websites, PDFs, YouTube videos, and other resources.
Test, Quiz and Assess
Learning is not just about acquiring knowledge. A significant part is also to show what you know. Demonstrating this knowledge is crucial as it allows for feedback, improvement, confirmation of your knowledge to others, and sometimes, it’s a pathway to earning degrees or certificates.
This is especially true in workplace settings, such as during compliance training, where one might need to prove to an auditor that the team has been instructed about anti-money laundering regulation. Whether you validate that customer service agents are product experts, or you acknowledge your sales reps with a badge, you are demonstrating the value of assessment. In tasks such as testing, quizzing, and assessing, AI can truly unleash its potential. This is what you can do:
General knowledge
When you use ChatGPT, Bard or Perplexity, you can prompt the AI to assess a certain input against the criteria you define. The AI will formulate its response based on the knowledge it already has (with all its strengths and weaknesses; see above). Example prompts:
- check the following paragraph for spelling errors
- tell me which part of my answer is correct and what needs improvement
- create a quiz about the French Revolution
Specialized knowledge
When addressing specialized knowledge, particularly in employee training, it’s crucial to integrate relevant source content into your prompts. This provides the AI with a foundation for its evaluations. For instance, you might copy and paste a segment from your product fact sheet or code of conduct into ChatGPT, instructing it to use this information as the basis for its responses.
A word about data
Should we really copy our internal content into ChatGPT or Bard? you might wonder. Good question. Yes, if you want to use these AI-Chatbots directly and want to build on internal knowledge, you need to give them the material to work with. It just doesn’t have access to your internal knowledge base.
OpenAI, the provider of ChatGPT, for example, can use your interactions to further train their Large Language Models. However, there’s an option to opt out, with the side effect that you won’t have access to the chat history. Another option is to purchase their Enterprise version.
When you use a specialized tool with AI integration, it employs AI through a dedicated API (a set of tools and protocols that allows different software applications to communicate with each other). In the case of OpenAI, they have declared that they won’t use the data received through their API to train their models and will delete it after 30 days. Therefore, using a specialized tool like eggheads is advantageous from a data privacy and security standpoint.
Specialized Tools
The market is already populated with various software programs designed for testing, assessing, or quizzing, each with its own set of capabilities. For example, let’s look at how you can use eggheads as an assessment tool and as an AI quiz maker that builds on your material.
eggheads as an AI Quiz Maker
To automatically generate a quiz with AI, you first enter your source content into eggheads. Then, you choose the type of quiz you want to generate, such as a multiple-choice quiz, a true or false quiz, or a fill-in-the-gap quiz. Select the length and language of the quiz, and in a few moments, you have a chat-based quiz that you can share via a link or Microsoft Teams.
You can, of course, generate the quiz directly using ChatGPT if you provide the right prompt. However, from there, you would have to transfer it to another tool to make it available for users and track it. If you need a format that you can share and wish to gather data on how it’s being used, eggheads is the tool to consider. It readily provides sharing options and deeper insights into the knowledge level of your audience through analytics.
eggheads as an AI tutor for Assessments and Personal Feedback
We’ve already discussed the potential of AI to become everyone’s personal tutor, providing feedback for improvement. Eggheads offers the capability to create a customized tutor for your company. You simply submit information relevant to your organization that eggheads’ AI will use as the basis for providing feedback. Users can be asked open-ended questions, and the AI will deliver personalized feedback, highlighting areas where they’ve performed well and suggesting areas for improvement.
In employee training, you can apply eggheads to assess knowledge or simulate role-play scenarios. For instance, sales reps could be asked how they would counter a potential customer’s objection, customer service agents might be asked how they would handle an angry customer, or employees could be quizzed on how they would respond to a phone request to transfer a large sum of money.
By choosing the right tools and using them responsibly, AI can significantly enhance the learning experience, making it more interactive, personalized, and effective.
In conclusion, with a focus on coherent flow, specificity, ethical considerations, and clear definitions, AI can be harnessed effectively and ethically for learning and assessment purposes, benefiting both learners and educators.
4. ChatGPT as your Sparring Partner
ChatGPT and other Large Language Models contain not just the knowledge found in their training data but also the most widespread perspectives and opinions. They mirror the patterns, biases, and viewpoints they learned during their training.
This makes them really useful, especially when you want to understand views and ideas that are different from your own. Or when you need another voice or perspective to help you overcome an obstacle and keep going.
Change of perspective
In our article about writing prompts, we mentioned that giving ChatGPT a role is a key component of a good prompt. This sets the right context and leads to better-targeted outcomes.
It’s quite similar when you use ChatGPT to get another perspective, as it shifts the lens through which you perceive a problem.
Your participants
In employee training, better understanding your participants and stakeholders is key to creating more relevant training. For example, when you’re designing a course about data privacy for marketing but you understand little about marketing, take advantage of AI. You can ask ChatGPT about the role of a marketer, the major pain points they face, their hopes and fears, and their perception of data privacy. This will provide you with valuable insights about the participants of your training, enabling you to tailor the messages and content accordingly.
Your stakeholders
When engaging with stakeholders, whether they are subject matter experts in employee training or parents of school children, ChatGPT can uncover what someone in their position might think. What might their expectations be? What might they do in this situation?
ChatGPT won’t give you the exact answers, but material for you to consider and work with. Simply put, it helps you put yourself in someone else’s shoes.
Generate and Brainstorm Ideas
When faced with a challenge, you might have some initial ideas. Yet, if you hit a roadblock and ideas stop flowing, ChatGPT can help get things moving again. Use this AI to dig deeper into the topic, find new connections, and essentially broaden your horizon.
Brainstorming with ChatGPT gives you many more options to choose from. If the number of new ideas seems too much, you can use ChatGPT to help sort through the options based on your own criteria. Staring at a blank page can be scary, but with ChatGPT, you have a brainstorming buddy ready to help.
Examples of using AI for brainstorming include generating ideas for your team event or school trip, or coming up with warm-up exercises.
Get feedback
Four eyes are better than two. Having a coworker read through a draft of your script or review your course outline can provide valuable feedback that improves the quality of your work. Similarly, ChatGPT can review the material you present to it. Ask it to review it, point out what’s working and what’s not, or suggest how it could be improved. The feedback you get is likely to highlight certain aspects you weren’t aware of before.
If you’re up for a challenge, you can direct ChatGPT to provide more critical feedback with prompts like “be a harsh critic” for those who can handle tough feedback, or “Tell me why this won’t work” to identify potential blind spots. For a specific perspective, use prompts like “Judge from an xy point of view” or “Give arguments for and against” to explore different angles.
Get advice – ChatGPT as your AI Coach and Council
Whether you’re a one-woman L&D department or a solopreneur, work can sometimes feel quite lonely. A bit of motivation and advice from a coach can provide the nudge needed to keep moving forward. However, not everyone has access to such guidance. ChatGPT can fill this gap by offering motivation and advice when you need it, acting as a virtual coach and council.
Put ChatGPT in Your Position
For example, you can ask ChatGPT for advice on your teaching or instructional design method. It can provide feedback on what you’ve already created. But, you can also ask it for guidance by saying: ‘What would you do, ChatGPT? How would you solve this? What’s the next step?
Put ChatGPT in Someone Else’s Position
You can ask ChatGPT to think like someone else, like a famous person, and give you advice. For example, have it embody Steve Jobs to coach you on your learning innovation project or edtech idea or ask what they would do if they were you.
Debate and Role-Playing
Engaging in debates or role-playing exercises can help you see things from different angles and fine-tune your message. This is really useful when you’re getting ready to talk to stakeholders, share your ideas with them, or meet them in person. Understanding different viewpoints is helpful, and debating with ChatGPT can help with that.
Between you and ChatGPT
You can use ChatGPT to practice these interactions, hone your arguments, and build your confidence. For example, you might pretend ChatGPT is a CEO and discuss your new onboarding program with it. You could also bring up recent topics from the last town hall meeting to make the discussion more relevant.
Between (historical) characters
For a fun exercise, you could have ChatGPT simulate a debate between historical figures, like Steve Jobs and Aristotle, on a topic like innovation. This is how it might go:
Aristotle: Good evening. I shall begin by positing that innovation should not be pursued recklessly. It should be guided by a foundation of established knowledge and virtue. Our quest for the novel must not overshadow the time-tested truths that have sustained civilizations.
Steve Jobs: Good evening. While I respect the foundations, I believe that innovation is about challenging the status quo and breaking the old molds to make way for the new. It’s through this process that we evolve and improve our conditions.
Wrap up
AI like ChatGPT can offer new viewpoints, help improve ideas, give helpful feedback, and more. ChatGPT is a handy tool for both creative and work tasks. It’s not about taking away human contact, but making you better prepared and helping you learn more.