Marcin FrąckiewiczFor decades, people have wondered about the potential of AI. With the technology
right in our hands, people are asking: Is AI really that smart? Let’s find out!
“AI as a tool in music-making is fine, but it’s always going to be the humanity in music that makes people want to listen to it.” – Jacob Collier
Seems like Artificial Intelligence has been all the rage in the past 2 years. Similarly, everyone was talking about how Cryptocurrencies are going to end fiat money once and for all.
So, is there any merit to all these claims of AI being as smart as people say it is? Or is it just another tech fad that’s going to die down in the next few months?
The rise in popularity of ChatGPT in 2023 caused a lot of different reactions from the public. Some touted smart AI as a useful tool with the potential to change how we do things today. Others have been skeptical and think AI is going to go rogue and take over humankind.
However, we are here with a different mindset and we’re going to look at where the current AI technology stands and how it may change in the future.
Without further ado, let’s dive right in.
- The History of Artificial Intelligence
- Can AI Think for Itself?
- Artificial Systems: Mimicry vs. Understanding
- The Future of AI as Partners, Not Replacements
The History of Artificial Intelligence
The concept of AI isn’t new at all. It’s just that it has come into the limelight in the past few years. In fact, the concept has existed for more than a century.
It all started in 1950 when Alan Turing published Computer Machinery and Intelligence. This proposed a method to test a machine’s intelligence using a test called The Imitation Game, now known as the Turing Test.
By the late 60s, Joseph Weizenbaum had created a chatterbot, which we now know as chatbots. It used Natural Language Processing (NLP) to talk to humans. Around this time, Alexey Ivakhnenko proposed an approach to Artificial Intelligence which we now know as Deep Learning.
During the 80s, the AI boom had already started. Japan’s government invested more than $2 Billion in today’s money to create computers that could talk to humans in a natural language. Autonomous vehicles were developed, and software that uses AI was created.
Around the 2000s, funding and interest in AI decreased significantly which led to a slower development. Still, during this time, Microsoft, IBM, NASA, and Apple, as well as companies like Netflix, Facebook, and Twitter started using AI to improve user experience.
After 2010, AI made a lot of progress. Researchers from Google trained neural networks to recognize cats by just using an image. A robot named Sophia was developed by Hanson Robotics which had a realistic appearance as well as ways to communicate emotions.
All of these developments led us to 2020 when OpenAI started testing GPT-3, which used Deep Learning models to write, create poetry, and develop code.
Can AI Think for Itself?
When we first got access to ChatGPT, it was a magical experience. It felt like there was a real human sitting on the other side of the screen, which was an astonishing, but also scary experience.
Currently, AI is mostly limited to the cloud, since most everyday devices aren’t powerful enough to power the calculations needed for AI. However, with a fast and reliable provider like Xfinity Internet, you can access AI features even on underpowered smartphones and PCs.
This had us wondering if there’s going to be a point where AI will be able to think for itself.
Before we talk about other things, let’s differentiate between knowledge and intelligence.
Knowledge is a collection of information and facts whereas intelligence defines the ability to use that knowledge to analyze and solve problems.
Take ChatGPT as an example. It was made to mimic human patterns during conversation. It can also give you answers based on knowledge it has or from the internet. However, it can only hold a conversation for a few minutes before losing the whole context .
Meanwhile, human intelligence is better at understanding the world around it and making decisions, even when entirely new situations arise. Human intelligence is multi-dimensional and complex.
So, to answer the main question – No, AI can’t reason, think, or remember things the way humans can. Well, not right now at least.
AI can use training data to learn whereas humans learn from every single interaction which means that AI learns in steps and humans learn exponentially. Even with this, AI can’t form memories yet or draw conclusions from experiences as well as humans can.
Keeping this in mind, AI is only as good as the data it’s given.
Artificial Systems: Mimicry vs. Understanding
Remember when AI image creators like Midjourney blew up? While the art they created was impressive, to say the least, there was a huge issue with them.
You see, the model was trained on millions of images from other artists who didn’t consent for their art to be used for training. This is why, images generated with Midjourney had scribbles in their corners. These were the artist’s signatures in the corner which are usually done for the artist to maintain ownership.
This proved one thing. The AI was mimicking art, rather than creating it.
This makes it hard to argue that pre-trained models actually “understand” language the way humans do. But, that’s what makes Artificial Intelligence “artificial”.
However, the newer GPT-4o model by OpenAI is a great example of how AI is moving in the direction of being able to understand what the user says and respond accordingly. It seems to have the ability to connect dots in a sea of information while making sense of it all.
The Future of AI as Partners, Not Replacements
While AI can be “smart” to some extent, the way we look at AI needs to change. Most people think that AI is out there to replace us. However, the future of AI is collaboration rather than replacement.
Right now, there are a lot of uses of AI that involve doing mundane tasks, which can free us to explore newer ideas and solve bigger problems.
Rather than viewing AI as a threat to our jobs and creativity, it’s a powerful ally that can help us be more efficient, innovative, and save time.
To answer the title’s question, AI isn’t that smart yet and it certainly isn’t a fad that’s going to die down in a few years.
The future seems bright and it seems like we’re only getting started.
