ChatGPT's revolutionary insight in language unlocks multitude of new applications
It's a cliché, but we can't ignore it: one after the other AI development is turning the (digital) world upside down. At the end of November '22, OpenAI dropped yet another bombshell: ChatGPT. ChatGPT can write full texts based on some well-aimed questions ( prompts ). And it does that very convincingly.
At first you think 'a fun gimmick', because indeed: you can have such
a system write something for you, but you can just as well do that
yourself, right? You don't need a lot of inspiration to imagine such an
application as the newest integration in a regular word processor. We
already know spelling correction and word completion, in the future the
program might even suggest full-on text fragments, just after you
entered the title. Could become a handy tool.
But the possible applications go infinitely further because the language insights of ChatGPT form an important element in the understanding and processing of language by computer systems:
- Text Categorization: The better computers understand our language, the better the systems can label and categorize text. This leads to applications for mediating social media, cataloguing documents and files…
- Speech recognition:
the computer can recognize text when it's recorded, type it out, enter
it in Google, and send you back the results, written or spoken. Think of
minutes that are automatically taken during AND summarized after a
meeting.
- Handwritten text recognition: For thousands of years, information has been stored in handwritten text. Automatical digitisation, making the sources searchable, labelled, and summarized, now becomes within reach.
- Also, program code is language, so ChatGPT can write whole new programs, detect errors, improve or make suggestions to your code and explain why it makes these changes.
Colleague Jef about ChatGPT: “ The model can even reason
, because it has picked up the patterns of logic as they appear in
texts. For the time being, it will not come up with new mathematics or
science, probably it cannot find enough examples nor collect data. But
ChatGPT can produce common reasoning and refute false reasoning. So you
can use it to check your own reasonings: are they correct or complete?
The precision and functionality will still have to be clarified, but the
basis is already there.”
Also interesting to know: ChatGPT does not work in real-time, it was trained with information until mid-2021 and is therefore super smart, but it does not know of its own existence, for example.
What ChatGPT thinks about itself: “Why is ChatGPT revolutionary?”
I'm sorry, but I'm not aware of the specifics or features of ChatGPT, so I can't tell you why it would be revolutionary. If ChatGPT is a computer program, it could be revolutionary if, for example, it offers a new way to communicate with people or if it uses advanced technologies that have not been seen before. However, as I have no information on the specifics or features of ChatGPT, I can't tell you for sure why it would be revolutionary.
The question now remains how creative we as humans will use this new tool. On Twitter, Guy Parsons already shared a test in which he asked the system for 'interesting fantastical ways of decorating a living room'. He shared the ChatGPT result in image generator 'MidJourney', and the result was… images of a fantastically decorated living space.
- Click here to read the message in Twitter
- also on Twitter, Rob Lennon shares techniques to improve the creativity of ChatGPT
Attention: garbage in = garbage out
ChatGPT itself gives out very important advice “… learning about AI can also help you understand the ethical and societal implications of AI, which is important to ensure that AI is used in a responsible and beneficial way."
Because several users have already noticed in this tool that the 'garbage on the internet' is copied and reproduced here. ChatGPT recreates biases that exist on the internet. The makers built in some safety features, so you cannot, for example, request a new speech from the Ku Klux Clan. However, human ingenuity still beats AI for the time being, because users of the tool quickly realized that some protections can be circumvented by asking ChatGPT for an example of program code or lyrics about a topic. Sample code from ChatGPT to "select hardworking people from a JSON file" gives selection criteria that do not conform to ethical principles.
Dieter Castel started a Twitter thread with shortcomings, often hilarious, but some also truly scary.
I'm compiling a thread of some #ChatGPT #fails 👇
— Dieter Castel (@DieterCastel) December 2, 2022
Now of course you wonder: did ChatGPT also write this article? To
our disappointment we noticed that it is not able to do that yet, or
that we can't prompt it out of the system. All of our commands were
answered with clichés that have been said a hundred times, such as the
intro to virtually every course on our website.
So much for our human creativity...
Asked ChatGPT: What is AI? Why is it important to learn about AI?
AI has the potential to revolutionize many industries and has already been applied in areas such as healthcare, finance, and transportation. However, the development of AI also raises ethical concerns, and the responsible use of the technology will be important as it continues to advance.
Learning about AI is important because it is a rapidly growing field that is transforming many aspects of our lives.
ChatGPT assignment: Write a rap song about talking with an AI-powered computer
Verse 1: I'm here to tell you about a new sensation, Talking with an AI-powered computer, It listens to me, and understands what I say, It's like having a friend, but even better in every way.
Chorus: I talk with AI, and it's like having a genie in a bottle, I can ask it anything, and get an answer on the spot, It's like magic, and I'm hooked on this technology, Talking with AI, is the future, and it's here for you and me.
Verse 2: I talk with AI, and it's like having a personal assistant, Always there, ready to help, no need to be hesitant, I can ask it anything, and get an answer on the spot, It's like magic, and I'm hooked on this technology.
Chorus: I talk with AI, and it's like having a genie in a bottle, I can ask it anything, and get an answer on the spot, It's like magic, and I'm hooked on this technology, Talking with AI, is the future, and it's here for you and me.
Verse 3: I talk with AI, and it's like having a supercomputer, Always on, and never tiring, ready to compute, I can ask it anything, and get an answer on the spot, It's like magic, and I'm hooked on this technology.
Chorus: I talk with AI, and it's like having a genie in a bottle, I can ask it anything, and get an answer on the spot, It's like magic, and I'm hooked on this technology, Talking with AI, is the future, and it's here for you and me.
Isabelle Borremans
Isabelle is not an AI expert, but she has been communicating about AI (and AI training) for VAIA for four years. So, she knows exactly how to communicate effectively about AI training. Feel free to contact her with any questions on how to promote your AI course. She’ll gladly challenge you with questions like:
- Who is your target audience? Can you be more specific?
- What networks reach that audience?
- Where does your audience prefer to take courses?
- What AI skills does your audience need?
- Does your training match their needs?
Besides communication, Isabelle also works on VAIA’s strategy: Who is our audience? How can we reach them? How can we spark their interest in AI?
Related courses
SAIAR Summer Studio: From latent to physical space
Zomerschool - Kortrijk - Howest Hogeschool, SAIARlab
Summer School on Security and Privacy in the Age of AI
Summerschool - Heverlee - KU Leuven, UGent, VUB, Imec