In the previous post, I explored how useful ChatGPT currently is for data visualisation. This was done through testing this AI tool on a number of tasks and on its information retrieval ability.
The results showed that ChatGPT still had some way to go on the theory and consulting side of things, as it still made mistakes and had limited knowledge on the subject. Despite this, it’s still impressive how it can easily provide a written response that was mostly correct for what I had asked.
Where ChatGPT really shined was on the coding side of DataViz. You can simply request the code for a particular chart in a number of different languages or libraries and ChatGPT will generate it for you. You can also ask it to update aspects of the code to change its appearance or functionality. ChatGPT can even correct code that isn’t working correctly and explain to you what you did wrong.
But if ChatGPT and other AI tools like it continue to improve, what impact will they have in the future for DataViz work? The ChatGPT interface already has the potential to revolutionise the way we retrieve information from the internet. So how disruptive would AI tools be to the field of data visualisation? Here I would like to speculate what the future might hold.
Negative Impact
First off, the main concern for people will be if any of these tools will negatively impact their lives. For the sake of being balanced, I would like to look at both the good and bad potential impacts of this emerging tech.
Potential Job Losses
Typically, it was believed that AI would completely replace blue-collar workers, but this hasn’t yet been the case. Robotics has been trickier to develop and before robot manual labourers are being built, we’re now seeing a bunch of new AI tools that can create art, write poems, compose music, and now can write code.
AI tools such as ChatGPT have demonstrated the potential to pose a threat to knowledge workers, yet the threat is not quite here yet. However, the fear is that if AI tools continue to develop in the direction it’s currently going in, a lot of white-collar workers will be at risk of losing their job.
For those unaware, last year an AI-generated picture won an art competition and many artists were unhappy about it. DALL-E and its offshoots can create realistic-looking images and art from a description in natural language.
There are also a number of AI tools out there that can aid in copywriting and writing texts such as Jasper, Writesonic, and Wordtune just to name a few. All you have to do is describe what you want to be written and these tools can generate anything from a block of text to an entire article.
Sound and music production aren’t safe either. Researchers at Google have been working on an AI that can generate minutes-long musical pieces from text prompts. OpenAI also has Jukebox that they’re developing.
So it’s not a massive leap of the imagination that sometime in the near future, someone will develop a tool that will do the work of an infographic designer or data journalist. If you look back at the previous post, ChatGPT is already providing some advice and information on data visualisation. While it still makes a lot of mistakes, if AI tools like ChatGPT continued to develop and absorb all the information on data visualisation from websites, blogs, and books, it could potentially fill the role of a consultant and educator.
This could also mean that training courses would become redundant. What’s the point of paying for an expensive training course or even spending many hours on a free MOOC course, when you can just ask an AI about a particular chart type, and how to replicate it in JavaScript or any DataViz application. The solutions would be in a matter of seconds and wouldn’t require much investment.
Perhaps designers who specialise in making highly custom and unique visualisations would be safe from the onslaught of AI. Despite this, I imagine them to be fewer in numbers as their work is for a more niche audience.
Resource Websites and Dataviz Tools Redundancy
AI could also completely change the way we consume information from the Internet. If an interface such as ChatGPT can simply provide answers and solution to any questions or request you ask it, then what would be the point of reading and searching through webpages?
Perhaps in the future, AI tools will reduce the importance of webpages as the way we acquire information will change. Maybe the popularity of things like articles, blog posts, brochure websites, and even Wikipedia pages will be greatly reduced.
Resource websites such as this would become redundant if an AI interface evolved to a point where could answer all your enquiries about different chart types. Also, many of the DataViz tools linked here that are used to generate visualisation could also become redundant (in their current form) with the advent of AI. ChatGPT can already write code for drawing charts, so there’s no need to refer back to a repository website for chart code.
If AI tools develop the ability to generate image files of a chart or create an interactive dashboard from an uploaded dataset, then there will be a threat to many web and desktop based DataViz drawing applications. Already, something like this is probably technically feasible, it’s just a matter of time before this capability is developed into an AI application.
A Deluge of Mediocrity
Although this isn’t specific to DataViz, the effect AI will have on the quality of content online could be incredibly negative. As AI tools make it quicker and easier to produce more content, we will likely see a surge in the amount of AI-generated content online. Perhaps AI content will even eventually overtake human-generated content online.
However, from what we’ve seen so far from AI, the quality of the content it produces ranges from bad to pretty good, but nothing exceptional. Artificial general intelligence (AGI) also known as strong AI, still seems far off. Our current AI technology can only learn and create content from knowledge and ideas that already exist. Making it only capable of churning out stale and uncreative content on mass. It lacks the ‘spark’ of the human mind to produce ingenuity and creativity. Therefore, if AI content flooded the internet, we could end up drowning in a deluge of mediocrity.
There’s been a conspiracy theory going around in the past few years known as the Dead Internet Theory. In short, it’s the theory that the original, “human” internet died around 2016-17 and has now been taken over by bots who produce content to manipulate people and to get them to buy stuff. While this theory may seem somewhat farfetched, it might become a reality in the near future with the development of AI technologies mass producing content online.
Fake Data Visualisation
In the realm of DataViz, we could see AI churning out loads of charts to push an agenda. In the previous post, we’ve seen how ChatGPT is capable of generating dummy data. So it’s not a stretch of the imagination that AI would be capable of mass-producing charts based on fake datasets. Even if you include a source at the bottom of the chart, that source itself could be entirely AI-generated as well and trick anyone not willing to investigate deeper.
If there were to be a serious epidemic of fake charts flooding the internet, then of course people would no longer trust charts they see online. How to combat this problem would be an issue we would have to deal with.
Impact on Design Decisions
Attention spans have decreased noticeably over the past decade or so. We can see this clearly with the rise in popularity of Tiktok and YouTube shorts, especially with the younger generations. If this trend (unfortunately I must say) continues, then it will no doubt impact the way people search for and consume data-focused content.
In this environment, how will charts and infographics be designed? How would it change the way we tell data stories?
One obvious answer is that content will have to be short, easy to grasp, and engaging enough the grab attention. So maybe small, clear charts that are animated. We’re already seeing that with some of the solutions Flourish provides such as the race bar chart and animated pie chart.
If AI tools develop DataViz capabilities, then a lot of visualisations will be generated within AI interfaces as well. How would tools like ChatGPT present data and tell data stories? Will there be built-in story templates that an AI tool will follow? Paolo Ciuccarelli, founder of DensityDesign research lab recently wrote a post on What AI knows about data visualization and data storytelling.
Of course, we don’t know exactly how the culture and mentality of people will be in the future. Maybe there will be a rejection of ultra-short-form content, short attention spans, and AI-generated content in general. As a response, there could be a return to appreciating more long-form content and authenticity.
Positive Impact
So far, most of this post has focused on the negative impact of AI, so let’s try to balance things out by being more optimistic and speculating about the positive impact on DataViz.
Increased Efficiency, Insights and Accessibility
While there will be some inevitable job losses that will not be completely replaced with new roles, AI will likely augment human abilities and make people more productive and efficient. Already, AI is being used in the gaming industry to generate game assets, saving time and money, while still delivering quality.
By utilising automation and machine learning algorithms, analysing massive datasets could be made a lot easier. Firstly, AI could be used to reduce the need for manual data cleaning and preparation, which would free up time for DataViz professionals. Secondly, AI could be used to identify patterns and insights that would normally be difficult for a human to detect due to the scale of such large datasets.
Additionally, AI tools could make it easier and more accessible for non-experts and those new to DataViz to understand and utilise data visualisations. In the previous post, ChatGPT demonstrated its ability to generate code for certain chart types, making it possible for those with not much coding ability to produce their own code generated charts. Therefore, if an AI-powered tool for data visualisation was developed, it would lower the barrier for more people to get into data visualisation work.
Amplify Human Intelligence with New Tools
AI can be used to enhance existing tools and to create new tools to amplify human abilities. In an essay on Distill by Shan Carter and Michael Nielsen, they describe and demonstrate how AI can be used to augment human intelligence. Two of the examples included show how AI tools can be used in the design of fonts and shoes by generating many variations which could be used as a source of inspiration.
For DataViz, these generative machine learning models could be adapted for rapid prototyping of visualisations or dashboard design. It might even be possible for AI to aid in the design of new forms of visualisations with these generative model tools.
Eventually, there will be an AI tool developed that specialise in data visualisation work. Such a tool wouldn’t do all the work for you but would be a DataViz AI assistant that could aid in all related tasks such as:
- Finding datasets online
- Data wrangling
- Analysis and insights
- Recommendations and guidance
- Generating visualisations and code
- Publishing in the form of dashboards, infographics, data stories templates, etc.
This might be a tool where you would upload your data and prompt the AI what you want to show or analyse and the format you want it presented in. The AI could then guide you through the process, and provide you with recommendations and various options to select.
Is It Just All Overblown Hype?
In the end, this is all just speculation, which can be fun to do. But I could be completely wrong about all this and this recent craze around AI interfaces like ChatGPT, which could be nothing more than hype. It will all depend on how this technology develops in the coming years and if the current issues can be solved with new innovative solutions.
Maybe AI will turn out to be a disappointment because the amount of improvement that comes about doesn’t meet the sky-high expectation generated from all the hype. Especially when there are dramatic claims of how it will revolutionise society.
However, I believe there’s a good chance that AI interfaces will have at least some impact on the way we as data visualisation professionals work. Especially in the tools we use to do our work. The key factor will be for us to adapt and acquire new skills in order to stay relevant in the industry. If AI becomes an integral part of our workflow process, then learning to write prompts effectively to get the outcomes you require will be an important skill.
As for this website, if it were to be completely updated in the AI age, I think it would have to be more than just an expanded chart list with more in-depth content. There will also need to be an AI component to it that will aid in searching, informing and teaching. But whether that is feasible and worthwhile is something to consider…
