We know what it does… But how ♂️ (ML focused for the most part)
As we all know, there has been an absolutely massive surge in interest and utilization of Artificial Intelligence Software (AI). We could blame ChatGPT for breaking the internet recently with Its crazy cool capabilities. But AI in general is a fascinating topic, featuring in conversations since 1956 (Among scientists), and further experiencing a huge spike in the 21st Century!
Even being the main focus of a few classic films such as The Matrix, The Terminator, The Iron Giant, and many more. Which either makes us paranoid for the future (In an entertaining way ) or wholesome…
But have you ever thought about how AI actually works? Keep reading to find out in a very basic way!
Table of Contents:
- What is AI ?
- How does AI even work !?
- The future of Artificial Intelligence …
- ChatGPT by OpenAI.
- Christmas Bonus: Asking AI If AI could end humanity
What is AI ?
As we all know humans have a certain level of intelligence allowing us to respond to specific questions, and learn new things, amongst many other things. Now, what If we could build/program this, and then give this capability to robots? That is AI in a nutshell.
AI is, “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.” — AI by B.J Copeland.
How does AI even work !?
AI works by merging huge data sets with different algorithms in order to discover patterns in the data that is processed. Each time a data set is computed (read by the computer) the AI is able to then monitor its own performance, and store this information for the future (kind of like how humans learn, but much, much quicker).
But, whereas humans need a break when they are learning (Or sleep If you love constantly learning), AI does not. Becoming very well-versed in the task It was programmed to do, very quickly.
Let’s say you are an AI, and I had given you 5 different articles on how to play a game. Each article will explain the task in different ways and will either miss or add Information. But… You are able to notice a pattern, a similarity, which is the rules of the game, no matter how It is explained. And you would be able to fill in the gaps of knowledge you have from one article to the next, predicting the missing Information. And off-course If I give you 100 articles (Data sets), you would be more accurate In these predictions and your patterns will get more precise!
The future of Artificial Intelligence …
At least one thing is for sure, we won’t have to worry any time soon about killer robots taking over the world… For this to be possible, AI would have to reach a Human-Level, referred to as Artificial General Intelligence.
And once this level is achieved (Which might not happen), who is to say the AI will boot up one day and kill us all, what If It decides to aid mankind, and Invents things we couldn’t even comprehend to evolve us as a civilization?
“When prompted with the question “Is AI an existential threat to humanity,” Matthew O’Brien, a robotics engineer from the Georgia Institute of Technology wrote on Metafact: “The long-sought goal of a ‘general AI’ is not on the horizon. We simply do not know how to make a general adaptable intelligence, and it’s unclear how much more progress is needed to get to that point”.
The facts of the matter are that machines generally operate how they’re programmed to and we are a long way from developing the ASI (artificial superintelligence) needed for this “takeover” to even be feasible.” — Could AI Wipe us out? The Sun
ChatGPT by OpenAI
In what felt like a few seconds ChatGPT took the industry by storm, everyone and their friends had heard about it and were trailing the different tasks the AI could perform. From Writing blogs, fabricating poems, and completing homework, all the way to debugging code snippets!
We can all see why ChatGPT is so popular. But how does ChatGPT learn new things? Not only was ChatGPT trained on the usual huge data sets from the internet, but also conversational data, and utilizing RLHF.
Another thing to note… The software Isn’t accurate all the time. It can create plausible-sounding but completely Illogical answers.
“Trained on huge amounts of data from the Internet, including conversations, it was also trained using a machine learning technique called Reinforcement Learning from Human Feedback (RLHF), in which human trainers provided the model with conversations in which they played both the AI chatbot and the user.
Because of this method of learning, ChatGPT answers can come across as natural-sounding and human-like. And the bot is not just parroting text it has learned. According to Professor Mike Sharples, a professor of Educational Technology at the Open University in the UK, OpenAI’s language model is “creating an internal representation, not just at the surface text, but of the ideas and concepts behind it.” — Luke Hurst on ChatGPT.
Bonus: Asking AI If AI could end humanity
Here is a quick bonus from me to you (my readers).
Where I will be asking ChatGPT (AI) : Will AI end humanity?
Check out my latest blog post here with the answer: Asking AI, If AI will end humanity, and other fun questions.