We all know the hype about ChatGPT, the website-based application developed by OpenAI that answers almost every question you can think about. But what could it be its potential to help presenters? Is there room for AI presentations at this current state-of-the-art?
This article will explore what ChatGPT represents for business professionals looking to create quality presentations. As an extra, you will get a case study using ChatGPT to generate ideas to power up a sales presentation. Let’s get started.
Table of Contents
- What is ChatGPT?
- What does it mean to use AI in presentations?
- GPT for Slides
- How to use ChatGPT in presentations to engage viewers more effectively
- Case Study: Creating a smartwatch sales presentation with the help of ChatGPT
- Best practices for using ChatGPT in presentations
- Alternatives to ChatGPT for using AI in presentations
- Complementary AI for Presentations
- A much-expected upgrade: GPT-4
- Closing thoughts
What is ChatGPT?
ChatGPT is a deep learning language model developed by OpenAI, which is capable of generating human-like text based on the input provided. The main features of ChatGPT include the following:
- Generative Pre-trained Transformer (GPT): ChatGPT is a cutting-edge, artificial intelligence-based text generation model that utilizes transformer architecture to create linguistically similar text to human writing. This technology has been pre-trained on an immense amount of linguistic data, allowing it to generate intricate and complex sentences with incredibly high semantic richness. Moreover, the output produced by ChatGPT would be 50% longer than usual as it can draw upon a wide range of diverse facts and details to expand upon the original input. This makes ChatGPT an incredibly useful tool for generating long-form written content with ease and fluency.
- Large-scale training: ChatGPT was trained on a massive dataset consisting of over 45 terabytes of text data. The model works by taking contextual cues from the query and utilizing its store of knowledge to craft an appropriate response. Additionally, ChatGPT can learn from user feedback, so it keeps improving over time, providing better results each time the system is used.
- Contextual understanding: This AI tool can model various conversational scenarios using its powerful language modeling and natural language processing capabilities. Therefore, it allows ChatGPT to be able to recognize complex structures of conversations such as discourse, dialogue acts, sentiment analysis, summarization, etc. The system uses large conversational datasets to learn and generate meaningful responses. By modeling contextual information from these datasets, ChatGPT can comprehend nuances in conversations more accurately than existing AI systems. Furthermore, it can perform many other tasks like text classification or text summarization with high accuracy.
- Multiple use cases: ChatGPT can be used for various natural language processing tasks, including question answering, text generation, dialogue generation, and text classification, among others. It is an ideal tool for creating content from a different perspective or brainstorming content ideas.
- Customization: We are speaking about a natural language processing (NLP) tool that can be customized and fine-tuned to the specific requirements of any user or domain. Retraining ChatGPT involves adjusting various parameters such as word choice, sentence structure, and even word meanings to maximize its capabilities for the intended task.
- API-based deployment: ChatGPT can be easily integrated into applications through its API, allowing developers to build custom applications that leverage the power of deep learning for text generation.
All-in-all, we talk about a versatile tool – still in development – that can help us for a large scale of purposes. Users ought to keep in mind that responses, in some contexts, are not accurate since the software may know about extremely specific topics. Hence, it is constantly learning, and users can correct wrong inputs.
What does it mean to use AI in presentations?
AI tools are used in many different industries, but for presenters, ChatGPT presentations go from the perspective of content generation. Let’s think about the main issues presenters face as “creativity blocks”:
- Not knowing how to break concepts into simple terms or easy-to-apply steps.
- Not knowing how to summarize large chunks of text.
- Getting overwhelmed at the time of picking color combinations.
- Selecting which content should be allocated per slide.
On top of that, we can add lacking presentation skills and feeling stuck when working with time-restrictive presentations as two extra creativity blockers. Although they can be overcome with patience, harnessing new skills, and sometimes asking for professional feedback, AI tools can bring a different perspective.
Think of slides AI powered by deep learning as having a coach next to you to help brainstorm ideas for your presentations. You can ask the coach to give you a basic outline of a topic, a selection of titles for an article or presentation, and even take a step further and hint hooks to elevate the performance of your presentation. Some other tools, which we won’t dwell on here, can help generate better graphics or analyze data you need to insert into your slides.
GPT for Slides
We are all familiar with ChatGPT, but in fact, this software implements what’s known as Generative Pre-trained Transformer, the language model developed by OpenAI. Therefore, ChatGPT isn’t the only gateway to use GPT for presentations.
The key to unleashing GPT-3 power in presentation design is to work with the now-released open-access API that allows developers to use GPT-3 in their applications. Multiple applications already count with integrations with GPT-3, such as Zapier, Copyleaks, Code Snippets AI, and Notion AI, but the one that truly caught our interest is MagicSlides.
Although this is not equivalent to ChatGPT for Google Slides, MagicSlides is a third-party plugin that allows us to work with the GPT-3 API in Google Slides. To use it, you must count on an OpenAI account to create an API key access to GPT-3. Working with a single Gmail account is recommended to avoid performance issues. If your Google Chrome profile has multiple accounts, create a new profile from scratch and work only with one Gmail account.
Then, the concept is quite simple: you give the plugin a topic, the count of slides to create, and that’s it. You can add extra info as an optional feature, like instructions for which tone to use, which keywords to list, etc. The outcome is a simple presentation structure in a solid color background; you can then customize from that first iteration to your preference, like changing the Google Slides theme color, for example.
How to use ChatGPT in presentations to engage viewers more effectively
Depending on your goals, there are multiple takes on how you can use ChatGPT to make your presentations more engaging. For starters, you can use this deep learning model tool to create AI-powered presentations in the format of interactive Q&As, allowing users to ask questions and visualize the responses live. If your presentation has a multicultural audience, ChatGPT offers the option to provide real-time language translation, which can be integrated via subtitles into the presentation template.
Storytelling is a powerful asset for presentations, and through ChatGPT, users can generate narratives from scratch or based on a set of prompts. These stories can be personalized by adding interests from the audience’s members or preferences, so the story remains relevant to the overall presentation. Another important feature to highlight is that we can generate character dialogues for our stories, making the presentation more immersive and entertaining for the audience. Also, ChatGPT can add contextual data in the shape of statistics or big data information, turning them into understandable pieces of information rather than complex concepts.
In terms of visuals, we can create captions and descriptions or generate alt-text for images, videos, and animations, making them accessible-ready.
This same take can create accurate virtual tours through ChatGPT, providing information, context, and insights on specific locations during a presentation.
Gamification is another area to exploit through AI-powered presentations, and we can add that element to our slides in the following formats:
- AI-powered quizzes: By integrating ChatGPT in presentations, we can create AI-power quizzes or trivia games as alternatives to Q&A sessions. This serves to test the audience’s understanding of the topics exposed.
- Scavenger hunts: We can create a presentation powered by ChatGPT in which the audience has to solve puzzles or find content based on the presentation information.
For sales purposes, ChatGPT can help business professionals to generate predictive analytics models to understand market trends or customer behavior, making an impact in sales presentations.
We cannot conclude this section without mentioning the ChatGPT integration procedure’s basics required for some of the features exposed here. As users, we must integrate ChatGPT into the presentation software as an API or a plugin. Then, the input recognition helps the deep learning model understand text input from the user or the audience. Thanks to its NLP (Natural Language Processing) algorithms, the information passed through the ChatGPT API is processed to generate an appropriate response and streamed through selected mediums.
Case Study: Creating a smartwatch sales presentation with the help of ChatGPT
When planning this article, the main objective was to test, from a presenter’s perspective, introducing ChatGPT to craft a presentation. Therefore, we devised a hypothetical case: A brand wishes to present its new smartwatch at a live event. The presentation has to be streamed, and there is a select live audience attending the event. This presentation follows the format of other tech presentations like new releases.
Which are our main objectives to achieve in this case study?
- Obtain an outline of what the presentation should be about.
- According to the “sales presentation for a smartwatch” requirements, obtain a list of elements to include per slide.
- Generate suggestions on how to engage the audience to buy the product.
- Get insights on which elements to avoid while performing the presentation.
- Repurpose the presentation’s content to another format (article, social media, infographic, etc.)
So, our case study starts after login into ChatGPT and giving detailed instructions: “Create an outline for a sales presentation about smartwatches.”
ChatGPT generated this presentation outline:
- Brief overview of the topic
- Purpose of the presentation
- Importance of smartwatches in today’s world
- Benefits of smartwatches
- Health & fitness tracking
- Communication and notifications
- Mobile payments
- Time management and productivity
- Key features of smartwatches
- Display and design
- Operating systems
- Battery life
- Water resistance
- Sensors and tracking capabilities
- Product demonstration
- Showcase different models of smartwatches
- Highlight unique features of each model
- Live demonstration of various functions and capabilities
- Customer testimonials
- Share experiences of existing customers
- Highlight the impact of using a smartwatch in their daily lives
- Summary of benefits and key features
- Call to action
- Q&A session
Interestingly enough, it’s time to cross-reference with what we, as presenters, intend to create out of the presentation.
Slide #1 – Introduction
As suggested, the ChatGPT introduction PPT will be handled through the first slide. Since the brand presenting this smartwatch is already established and has brand recognition, the purpose of the presentation shall track back to some key features of prior presentations. This implies if the brand has an iconic aesthetic used to start the presentation, that format should be preserved.
The point about the importance of smartwatches in today’s world is rather obvious given the kind of event; therefore, it will be replaced with some stats on how the brand has helped customers through the years to improve their life thanks to these devices.
After the introduction is made, it should be immediately followed by an agenda slide.
Slide #2 – Benefits of smartwatches
Instead of generalizing the benefits of smartwatches, this second slide will directly speak about the features to be introduced in the to-be-released unit. It has to mention updates on health and fitness tracking functions (i.e., sleep pattern recognition, medication reminders, blood oxygen, wrist temperature, etc.), which new integrations it offers for mobile payments, productivity tracking apps, and how the notification system is updated. For convenience, we can directly mention a lighter watch specification (in terms of size and weight) or include those elements in the third slide.
Slide #3 – Key features of smartwatches
Now is the time to display your product in front of an audience. Some high-key photos taken by professionals should appear, as well as animations with a 360-view of the model.
Next, we have to talk about the specs. Battery life, display size, and type, display resistance, waterproof/weather resistant capabilities. The new OS version to include in the unit, with the revisions listed from a prior version. And to conclude, a list of the sensors paired with this smartwatch and their expected functionality.
This point doesn’t have to be a slide. It can be a blank slide. A slide featuring the title “Product Demonstration” or a video. The purpose is to give room to the presenter to engage with the device, showing its functionality in real time to the audience. Cameras should focus on the watch in detail as the presenter emulates daily life activities using the smartwatch.
Slide #5 – Customer testimonials
This can also be displayed in video format, with happy customers interacting with priorly-released units, to conclude with powerful graphics showing “the future” – a.k.a. the new model released.
Slide #6 – Closing
This is where you can become creative. A good point would be to feature an infographic comparing the functions and key specs between models (to highlight the reasons why users should upgrade). Call-to-Actions are to be introduced here, using the element of excitement to create interest in buying the product.
What ChatGPT suggested
We ran the exercise of asking ChatGPT what would be content to list on these slides, and here is the output.
Slide 1: Introduction
- A captivating image related to smartwatches (which we added)
- A brief overview of the topic
- The purpose of the presentation
- A statement about the importance of smartwatches in today’s world
Slide 2: Benefits of smartwatches
- Benefits include convenience, health and fitness tracking, communication and notifications, mobile payments, and time management and productivity.
- An image or graphic to illustrate each benefit.
Slide 3: Key features of smartwatches
- A list of features includes display and design, operating systems, battery life, water resistance, sensors, and tracking capabilities.
- An image or graphic to illustrate each feature.
Slide 4: Product demonstration
- Images or graphics of different models of smartwatches.
- A brief description of each model, highlighting unique features.
- A live demonstration of various functions and capabilities.
Slide 5: Customer testimonials
- Quotes from existing customers about their experiences with smartwatches.
- Images or graphics to accompany the quotes.
- Highlight the impact of using a smartwatch in their daily lives.
Slide 6: Closing
- A summary of the benefits and key features of smartwatches.
- A call to action (such as encouraging the audience to purchase a smartwatch)
- A space for questions and answers.
As we can see, ChatGPT’s approach isn’t far from what we accommodate the presentation to be. It is pretty generalist despite mentioning the case in first person during the user input, but that can be expected from AI deep learning models. Reiterative answers in some sections can also happen, although not as noticeable as other alternatives to ChatGPT we shall mention later.
Engaging the audience with ChatGPT
Points #1 and #2 of our case study are done, so we move to get suggestions from ChatGPT to engage the audience about buying the brand’s smartwatches.
This AI tool presents us with six different hooks to engage our audience.
- Pain point focus: We have to identify a common pain or problem many people face, then show how our smartwatch solves that problem. For example, it uses the case of difficulties experienced in time management and how a smartwatch can solve that problem.
- Personalization: This hook focuses on how the smartwatch can be customized to fit each individual’s needs and preferences. We can mention changing the watch face, the app market, or using custom bands.
- Health and wellness: By highlighting the health and fitness tracking capabilities, you can show customers how to lead a healthier and more active lifestyle.
- Convenience: This hook points to using the smartwatch in daily activities such as making phone calls on the go, enabling mobile payments, or browsing notifications without needing to check your phone.
- Emotional connection: Using customer testimonials, you can create personal stories demonstrating a smartwatch’s positive impact on someone’s life.
- Highlighting unique features: This would be your competitive edge. Explain which features make your smartwatch stand out in the market.
Recommendations on presentation mistakes
An advantage of GPT language is to develop accurate recommendations based on the context showcased. For our case study, we asked ChatGPT if there were any mistakes to avoid when giving this presentation about the release of a new smartwatch.
In this case, ChatGPT presents us with seven mistakes that presenters commonly make on topics like this. Lack of preparation we can put aside, as no professional presenter would arrive at a product release presentation without proper practice and assets.
The second mistake is related to bombarding your audience with technical details, which can end up boring the spectators. Instead, it offers the alternative of shifting focus to the product’s benefits and how it can help solve real-life problems.
Ignoring the competition as a mistake is questionable. You don’t see Apple speaking about Samsung smartwatches in their product releases, nor Sony talking about Nintendo or Xbox when presenting their newest PlayStation unit. That should go more from a preparation side when researching what the competition did to catch the audience’s attention.
Not showing the product! This suggestion surprised us. Honestly, how can you come up with a presentation about a product release and not even show a photo of the product itself?
Failing to answer questions can happen if the presentation stretches long enough to conclude the event, but it is not a common mistake in professional technology presentation events.
Poor visuals and slides, yes, we concede it is a common issue experienced by presenters, especially if they are overly populated with text.
Finally, it lists not practicing the presentation, which we also put aside as we did with the first mistake.
Overall, we ended up with 2 valid suggestions to look for out of the seven suggested. We can always retry the output, but when we tested that, the suggestions were pretty much similar.
Repurposing content with ChatGPT
So what if we intend to use the information in this presentation for another medium, for example, a blog post on the brand’s website. We asked ChatGPT to give 10 title ideas for blog posts on smartwatches for a tech website.
- “The Top 10 Features of the Latest Smartwatches”
- “Why Smartwatches Are the Future of Personal Technology”
- “How to Choose the Best Smartwatch for Your Lifestyle” – OK
- “The Health Benefits of Owning a Smartwatch” – OK
- “The Most Innovative Smartwatches of the Year”
- “Smartwatches vs. Traditional Watches: Which is Right for You?”
- “Maximizing Your Productivity with a Smartwatch” – OK
- “How Smartwatches are Changing the Game for Fitness Tracking” – OK
- “The Convenience of Mobile Payments with a Smartwatch” – OK
- “The Impact of Smartwatches on the Wearable Technology Market”
Out of the ten titles suggested, half are a good fit for the content of our presentation, which, of course, require additional research.
Asking ChatGPT for extra recommendations
So far, the outputs generated by ChatGPT have been interesting from a content perspective. But let’s ask some more questions to test its true capabilities as AI content.
We start by asking, on that same thread, how long the presentation should be.
It gives a vague answer about sales presentations, recommending sticking it below the 30-minute mark. The section about “clear, concise, and engaging manner” is presented as a vice through this thread – we believe we counted about 7-8 times that same sentence repeated in the entire chat.
We move on to another question, this time asking ChatGPT which would be the best visual medium to present a smartwatch: a detailed infographic or a video.
Again, we see the “clear, concise, and engaging manner” phrase in the second paragraph, which is somewhat rephrased in the final paragraph of this output. The suggestion retrieved is that infographics can help sales presentations highlight key features and benefits while also clarifying to the audience what your competitive advantage is. It recommends using videos as a bridge for emotional connection with the audience since they can demonstrate the product in action.
Finally, we wanted some suggestions about the color scheme in the presentation slides to encourage customers to buy the product. Below is an image of the output generated.
If you check our article about color theory, especially in the section about color psychology, the suggestions are okay. The question lies in a generalist output rather than explicitly suggesting a color palette for this. To be honest, we can retrieve the same answer when asking Google for the same search query. But what if we truly push ChatGPT to answer what we intend?
We reformulated the question with the following statement: “My brand’s colors are yellow, navy blue, and white. Can you suggest exactly which colors I should use for the slides?” Keep in mind the second question was right below the output generated above. Well, here’s what happened.
ChatGPT delivered this message: “That model is currently overloaded with other requests. you can retry your request, or contact us through our help center at help.openai.com if the error persists.”
This is an error we experienced multiple times when asking complex questions. Sometimes ChatGPT remains idle trying to process the answer. Other times, it writes an output (even quality ones) to halt them, and triggers that same error message, deleting the output generated up to that point – a frustrating experience.
Best practices for using ChatGPT in presentations
Now it’s time to brief some tips to help presenters extract the potential of ChatGPT for presentations at its current state-of-the-art.
Avoid complex inputs
As we’ve seen above, asking a very detailed question can make ChatGPT crash. You can question whether it is your mistake or just the server overloading. A good practice for this is checking ChatGPT’s homepage, and if the message shown below is displayed, there’s a server outage.
This can be flagged as a partial outage triggered by requests going way higher than the system can handle. Those peaks in navigation are seen at certain hours, especially during the Pacific Time afternoon. A complete server outage can happen, and that’s notified through a different warning, saying the service will be up soon.
Don’t rely on the text generated by ChatGPT
Especially if your intent is to repurpose the presentation’s content for a blog post, be EXTRA cautious when using AI tools like ChatGPT. Google is flagging and penalizing websites that retort to AI content generation. This is explained by the necessity of providing users with the best quality content available for their search queries. Your website can endure a severe demotion in rankings triggered by ChatGPT content.
A good practice is to rephrase the content manually, meaning to actually read what ChatGPT generated and explain it in your own words. That way, ChatGPT becomes an idea generator rather than a content generator.
Take a look at repeated outputs
You can always regenerate the response if the output doesn’t satisfy your needs or rephrase the question to make it more specific. We tested the option of saying “do not include…” but it works randomly.
Any AI tool for presentation is not intended to become an instant solution, and we cannot assume that we can entirely create a presentation with ChatGPT as the only resource. This process requires patience, skills to give the proper instruction (to avoid those repeated outputs), and a good deal of copywriting skills.
Outlines and title generation
ChatGPT is an amazingly effective tool for creating presentation outlines, helping a common’s writer block. Title generation is accurate, although it falls into the generalist category and can be easily flagged as AI content. We advise you to double-check titles, rephrase, or simply use the output for brainstorming.
Learning a new skill/definition
Say you have to deliver a presentation about a concept you don’t clearly understand – which, under some situations, can perfectly happen. ChatGPT can help you learn that concept to basic professional competence, and you can ask as many questions as desired to get a broader picture of the topic. Think about it as a personal, free coach.
Of course, do not take the outputs as absolutely true, as some responses can be biased through the machine learning procedures applied. With dedication, you can train ChatGPT yourself and help the community get better responses.
Yes, as cutting-edge as it sounds, the natural language processing features of ChatGPT and its machine learning algorithms help users evaluate their emotions in different datasets. How can this process be accomplished? Let’s go through a basic overview.
- Data Collection: Feed ChatGPT with transcripts from speeches or customer reviews.
- Model Review: ChatGPT uses machine learning algorithms such as Support Vector Machines (SVM) or Recurrent Neural Networks (RNN) to develop the emotion recognition model.
- Prediction: After the system processes the information, the output is sent using predictions that can be repurposed into new information.
- Integration: With the new information generated, we can create graphs and charts to turn into visually-understandable data the emotion analysis generated.
This kind of feature is ideal for processing large chunks of data regarding customer reviews and finding common patterns of conflict (i.e., areas in which customer service can be improved).
Alternatives to ChatGPT for using AI in presentations
Despite being the best-advertised option for GPT language tools, presenters should check other AI alternatives for content generation.
In our experience, Jasper one of the best content generator based on AI available in the market. We tested it in multiple scenarios, like generating metas, creating a presentation structure, generating step-by-step solutions, title ideas, and more.
We can work from pre-made templates or directly use the powerful Boss Mode, which is intended for long-format text generation, and it crawls the internet to search for valid answers.
Our objections? Jasper can be reiterative with its answers, rephrasing or reusing the same sentences in the L length (which is the maximum) of text generation. Much like what happens with ChatGPT, some answers fall into the generalist side, meaning there’s room for improvement.
Frase is an AI tool for content generation to improve Google search ranks.
We can appreciate considerable differences in keyword management for content generation (Jasper only allows 3 keywords), maybe due to this tool’s primary intent, which is web content.
Users can generate presentation structure, content briefs, large pieces of content, detailed analysis, and more, and customize where the AI engine is extracting the information from (if SERP, People Also Ask, Questions, Headers, Topics, etc.)
For presenters, Frase can be used to generate speeches or repurpose presentations into quality blog articles.
Tome.app ppt is a new multimedia tool to generate AI-powered presentations by integrating ChatGPT, Dall-E, and other interesting features.
The free mode offered by Tome is pretty simple. You sign in and then are offered a set of pre-made templates to test. Or, you can hit the create button and create a presentation from scratch.
You then write what the Tome presentation is about and click on create. Options for selecting which type of illustrations you want to add are listed, and if you intend to create a presentation, an outline, or a story.
It is an interesting take on presentation crafting. However, some users can feel it is restrictive due to using only Dall-E-generated images instead of uploading their photos.
Finally, we have iA Presenter, a tool still in the beta stage with a waitlist to access its features.
The proposal is truly interesting: you introduce text that can be extracted from multiple sources, and then you designate which are the headlines (using the # in front of the text). Slides are separate from each other by adding a – – – (triple dash) on the text.
Images can be easily incorporated, and there are plenty of options to create visual cues, such as gradients, which are better audience magnets than common bullet points.
Complementary AI for Presentations
In this section, we wish to list some of the tools we have tested to speed up the content creation process from scratch. Yet when we consider the future of AI presentation as something vague, these tools have caused an uproar in terms of content development.
Can you imagine converting text to images with accurate results? Well, Jasper upgraded their content offer with a new plan: Jasper Art & Write. You can then access the new Art functionality, which takes input from users and generates images in various styles: Cartoon, Illustration, 3D Rendering, Pixel Art, and Retro, among others.
Users must include a basic image description (although the more detailed, the better), select the artistic style, medium (like watercolor, acrylics, ink, etc. – much like traditional art mediums), artist/style, mood, and keywords.
Artist style surprised us with options like Dr. Seuss, Picasso, Pixar, and Van Gogh. It is a limited selection – 8 artists – with popular options like the South Park animation style.
Although you can generate images resembling stock photos, we wouldn’t advise using them in your presentations as they look quite odd when zoomed in. These limitations are likely to be lifted in upcoming revisions, but the current state-of-the-art does not encourage presenters to come up with photorealistic people results.
DALL-E 2 is the AI text-to-image generator used by multiple AI slides apps. It works similarly to JasperArt, but its integration with several applications helps presenters not to hop between software options.
The output generated by DALL-E is mostly geared toward artistic-style images, although you can come up with results pretty close to photorealism in landscapes and architectural images. The main advantage? Completely original images ready to be used in any presentation, blog post, social media, or video.
As the final AI image generation tool on this list, we should mention MidJourney, an AI tool that has raised controversies when winning an artist competition against real-life artists.
To access MidJourney, you have to apply to the official waiting list. After you are accepted, you’ll receive an invitation to their Discord server – a requirement to interact with the tool.
Like any Discord server, your interactions with the tool are public unless you create a private channel and invite the MidJourney bot to it. This would be our personal suggestion so you preserve the privacy of your projects. MidJourney works with credits in a subscription-based plan (basic tiers allow 200 images per month). The time to process each image depends on demand.
As we have seen with Jasper, DeepAI is a text-generator AI tool based on NLP (natural language processing). Its application is to create creative content as if written by humans, but we found its most powerful application in text translation.
If you used to struggle with making subtitles for your presentations, then take a look at DeepAI now.
We can all agree that video content production is expensive. Especially if we have to hire actors to do voiceovers or have them appear in video presentations. Synthesia.io takes video content creation a step further, allowing us to produce realistic videos with realistic AI avatars. These avatars are gender diverse, and they can fit multiple contexts, like doctors, corporate workers, athletes, common people, and even Santa Claus!
There is multiple language support available for Synthesia, with access to distinctive accents (like you want your avatar to talk in Mexican Spanish rather than Castillan). 50 customizable video templates are available for users, and you can also retrieve graphic assets from your computer or royalty-free services like Unsplash or Pexels.
This tool is subscription based, but you can start with a free trial. For a text-to-video tool, the fee doesn’t sound that expensive, especially considering the outcome and what it would be to produce such reels with real-life actors. We want to highlight that, in some cases, avatars look a bit robotic (like not blinking when speaking), so test the different options available with enough time to craft the best quality video for your needs.
If you want to deliver presentations with realistic AI voiceovers, then you ought to test Murf AI. It is a text-to-audio AI generator, with a library of more than 100 professional voice actors in 20 different languages. Creating voiceovers in Google Slides presentations is an easier task now as you can export voice, voice + music, and even subtitles in formats like MP3, WAV, or FLAC. Exporting to video is another option, but we find tools like Synthesia to be more powerful for this purpose.
A much-expected upgrade: GPT-4
With the announcement of GPT-4 by OpenAI, there is much guessing about what it will bring. We know that its application will roll only for ChatGPT Plus services, whereas the popular free service sticks to GPT-3.5 for the time being.
Given the reception of ChatGPT, and the confirmation of an API in the GPT-4 version, the new GPT release shall cover these main aspects:
- Visual content creation (images, videos, audio recordings from input).
- Content writing and educational material (apps like Duolingo and Khan Academy confirmed their integration with GPT-4).
- Better context understanding (something that GPT-3 had to be trained to grasp the user’s requirements properly).
So, in general lines, it may seem like a LOT of money to invest in subscription services, especially if we add to the mix the fees to access software like Microsoft 365, Adobe Suite, Canva, learning apps, entertainment, and the list goes on.
Who will benefit from using GPT-4? This is a good question considering how OpenAI targeted its marketing campaigns: 25k word processing limit from user’s input, web-crawling for text interaction, as well as improved conversations.
Companies and professional content creators are the number one target user for this GPT-4 version. The whole idea of the upgrade in the language model by OpenAI is to serve the needs of those who have continuous practical usage of AI tools to speed up their job, not casual hobbyists. And its application is expected to break all concepts considering how the U.S. government restricted access to China and Russia to the newer AI chips from Nvidia and AMD graphic cards.
Well, how does this premium version of GPT helps us, presenters? GPT-4 for presentations can deliver proper content structure to explain complex concepts, which GPT-3 and GPT-3.5 often struggled with if queries were on the rise. Working with big data and exporting information as reports is among the expectations of academic users. Also, given the experience with GPT-3, we believe GPT-4 will raise a vital asset to create engaging video content in presentations based on the audience’s interests and the topic to present. Again, these are predictions; we will confirm these hypotheses in the upcoming months.
The current state-of-the-art for AI presentations is rather incipient. 2022 can be taken as the first year in which AI started to fill the conversation topics and proved itself as a powerful variant to human-created content, although not a replacement.
We consider there’s room for improvement in ChatGPT for PowerPoint presentations and in all tools presented here, which naturally shall follow through usage. Regarding ChatGPT, the potential lies in idea generation, not in copying/pasting extracted content. Think of ChatGPT as an extra tool in your arsenal to become a better presenter, not the panacea for presentation design.