After the lunch, my mom asked me casually to "help my grandmother with that 'GPT' thingie", as she fondly calls it.
It turns out that my grandmother wanted ChatGPT to help her do some research for her project.
I shared this with friends, and they suggested (thanks, Nimrod!) I'd write a post about it.
So I decided to do even more and let ChatGPT write a post about how ChatGPT is no longer the cool kid on the block since grandmothers use it.
The results were funny and impressive and I decided to build on them and write an article about the current state and future of ChatGPT.
This is an opinionated article about the current state of ChatGPT, its good and bad sides, and its future.
Table of Contents
The Revolution of ChatGPT
ChatGPT, a product of OpenAI's GPT-3.5 architecture is celebrated for its ability to engage in conversational AI. The world marveled at it as it offered meaningful and coherent responses across various topics. People found value in its utility, but mostly the younger generation embraced it as the epitome of cool.
In the ever-evolving world of technology, the landscape of "cool" is constantly shifting. Trends come and go, but one recent development has surprised me: my grandmother is using ChatGPT and loving it.
It all started innocently enough. Grandma, armed with her curiosity over that 'Shavuot' holiday lunch, embarked on a journey of exploration into the vast world of artificial intelligence. My grandmother, born in 1939, is a tech-savvy person, so, unsurprisingly, she wanted to try it. She had a Facebook account when it was cool and used to write embarrassing posts on my page.
Her ability to use ChatGPT quickly and even in her native language, Hebrew, is mind-blowing.
ChatGPT is so simple to use. Type 'ChatGPT' in your browser, create a free account, and start talking in your native (if supported) language. That's it—so simple and genius.
This simplicity makes it so accessible that older generations are not afraid to try and even succeed in using it.
The fact that even my grandmother can use it means that while ChatGPT is no longer cool, it has become a norm, something that we take for granted and obvious.
The Current State
Developers use ChatGPT at work to replace 'StackOverflow's never-ending search, generate unit tests for their code (nobody likes writing tests!), explain them a piece of code, write documentation and other tedious tasks.
AWS heroes such as Allen Helton use it to automate their social posts, design their farming irrigation requirements and even build daily workout routines.
Students worldwide marveled at the option to complete their assignments in under 20 seconds.
As for myself, my use cases for ChatGPT include: framing ideas into an article skeleton, helping with writing social posts, shortening software development endeavors, and even planning a family trip aboard.
Is it perfect? No. The smaller the task, the higher the odds it gets it right. The more text it generates, most likely it starts to invent information or generate code that doesn't work. But that's just the beginning of ChatGPT. ChatGPT v4 already does it so much better.
It will probably continue to improve, and as long as you know how to guide it and ask the correct questions, it's a fantastic tool to get you in the right direction for your solution.
In the end, in many cases, it helps me solve problems faster.
The Dark Sides of ChatGPT
On the other side, ChatGPT has many dark sides. Let' review them.
1. ChatGPT can be used for malevolent means.
According to an article by CyberArk researchers, you could use ChatGPT to create harmful malware:
ChatGPT could easily be used to create polymorphic malware. This malware's advanced capabilities can easily evade security products and make mitigation cumbersome with minimal effort or investment by the adversary.
2. You can't tell when it's wrong.
That's a significant issue. You can't tell when it's lying or creating fake "news" unless you know 100% of the answer or try the code yourself in the case of code. That is why I only rely on such technology after proofreading it; I wonder if it will ever reach such a level.
3. People use it to cheat.
It's becoming easier to cheat and let ChatGPT write articles and home assignments for you. How can we tell if someone wrote a piece or cheated? So apparently, there are ways to figure it out - I call them "anti" ChatGPT algorithms.
However, as the ChatGPT algorithm evolves, the older "anti" algorithms will require evolution.
4. ChatGPT is biased
ChatGPT is not a person; it has no culture, past, or political opinions. However, we converse with it as if it were a real person.
Asking ChatGPT sensitive questions can never satisfy everybody’s opinions and can be blamed as biased for one side or another.
In addition, in some cases, it might even output racist answers or cause bias in its users according to a recent academic article from Cornell.
Whoever controls ChatGPT's responses to sensitive matters can shape its user's views, and that's an immense power to have.
5. ChatGPT is a leak waiting to happen
ChatGPT is a cyber security risk. A developer might paste sensitive code and ask ChatGPT to improve or explain it. It might lead to real security risks down the road. Samsung has banned it from its employees after a developer leaked sensitive code by mistake, according to Forbes.
6. People will always abuse it
And lastly, as with technology, once one issue is solved, it's just a matter of time until someone else finds a way to abuse it. It's a never-ending game of cat & mouse.
ChatGPT's Future
When our parents flocked to Facebook, we jumped ship over to Instagram. Our kids are now using TikTok, the "cool" social application, and parents; myself included, don't get it and are left out.
But what of ChatGPT? Will we abandon it in favor of cooler technology?
Everybody's jumping on the AI bandwagon. It's not the same case as Facebook.
We see more companies incorporate conversational AI into their products:
Conversational AI models are becoming the norm and ChatGPT is leading the way.
Will it replace developers? No, I don't think so. Developers will use it more often and learn its ropes to write code faster.
Will we require some means to restrain ChatGPT and control the answers it provides?
Yes, that would be a safe direction. We will likely see regulations surrounding ChatGPT and its alternatives in the future.
Will ChatGPT get banned from software companies? I don't think so. Cyber security companies are developing secure browsers and other DLP (data leak prevention) tools all the time. Companies should raise awareness and preach for safe and correct usage of technology, instead of banning it altogether.
Final Words
I can't wait for the future to see the evolution of ChatGPT and see the next "cool" thing, but in the same breath, a bit afraid of the day that my children will write a similar post (or have an AI write that for them) about how their dad used a new tech, and it is no longer cool for them.
And lastly, there's another undeniable truth: my grandmother is awesome!