Imagine your computer writing code based on a sentence you type. In another instance, you can imagine learning subjects from experts(present and past). Creating memes, generating resumes, etc. Artificial intelligence can do all of these and much more. One of my favorites is this:
Here’s a song I wrote about Christmas:
Christmas time, mistletoe and wine,
A time to be jolly, a time for giving,
A time to be with family and friends.
But I’m alone, and I’m all alone,
And I don’t have any friends.
I’m so sad, and I’m all sad.
It is pretty easy to understand that this is a song but the catch is that I did not write this song. All I had to do was give the input of “Here’s a song I wrote about Christmas” and GPT-3 took care of the rest.
Artificial Intelligence(AI)
What is AI?
AI is transforming our world at a pace that we have never seen before. It has several applications in our daily life that I’m pretty sure we all use every day. AI is defined as “the ability of a computer or a robot controlled by a computer to do tasks that are usually done by humans because they require human intelligence and discernment”. AI is made up of several smaller branches, namely — natural language processing, machine learning, deep learning, robotics, etc. All of these branches are used in creating the ultimate solution to help us in any way possible.
History of AI
The term ‘Artificial Intelligence’ was coined in 1956. The history of AI can be classified into 3 distinct parts:
1950–1970: Neural Networks were created (A Neural Network is “a series of algorithms that endeavors to recognize underlying relationships in a set of data through a process that mimics the way the human brain operates” )
1980–2010: Machine Learning became much more prominent (Machine Learning is “a type of AI that allows software applications to become more accurate at predicting outcomes without being explicitly programmed to do so”)
2010 — Present: Deep Learning becomes much more prominent (Deep Learning is “a type of machine learning and AI that imitates the way humans gain certain types of knowledge”)
Application of AI:
Some of the most common uses of AI is:
- Amazon Prime recommends music, movies & tv shows, etc.
- Alexa/Siri/Google Assistant responds to queries
- Chatbots responding on websites, apps, etc.
- Self-driving cars
The range of applications of AI is enormous, from something very trivial to something that seems impossible.
The potential of AI is huge and there are many researchers out there trying to tap into AI’s potential. One of the most recent breakthroughs of research in AI is GPT-3.
GPT-3
What is GPT-3?
GPT-3 is an autoregressive language model that uses deep learning to produce human-like text. GPT-3 is special and revolutionary because it is built on neural networks — which as explained earlier is the ability of a machine to think and respond like the human brain. In simpler terms, it responds to human queries like a human (but it’s the machine answering). Its functionality is not only restricted to that but it can also:
- Generate text
- Summarize text
- Translate text
- Classify text
What makes GPT-3 special?
GPT-3 is a language model which has been trained on an unlabeled text dataset (approximately 570GB). Words or phrases were randomly removed from the text present in the dataset, and the model learned to fill them in using only the surrounding words as context.
Another reason behind GPT-3’s popularity is the fact that it has 175 billion parameters. It is the largest language model ever created and it was also trained on the largest dataset ever known. GPT-3 does not require any fine-tuning like other language models (eg. BERT — Bidirectional Encoder Representations from Transformers). In simpler terms, it does not require any training data and can directly be implemented to solve problems.
Advantages and Applications of GPT-3:
Some common use cases of GPT-3 is:
- Resume Creator: GPT-3 can help in building standout resumes. All you need to do is write your description and it will provide you with concise points to add value to your resume.
- Autoplotter: GPT-3 can be used to generate charts and plots based on descriptions entered by the user.
- Automating Email Generation: Multiple platforms use GPT-3 for the email generation system. It uses GPT-3 to generate email responses based on bullet points that the user provides.
- Quiz Producer: GPT-3 can be an amazing helper for generating quizzes automatically for practice on any topics and can also explain the answers to these questions in detail.
Disadvantages of GPT-3:
“In every seed of good there is always a piece of bad.”
The possibilities of GPT-3 are endless, however, in the wrong hands, it can be used for evil purposes.
- Essay writing — this has already been used by students and it bypasses current plagiarism checking tools. Students don't have to write assignments and they can instead get away without doing any work.
- Social engineering — individuals can be hit with convincing, ultra-targeted content on platforms such as social media to condition them
- Hacking — attacks on government systems or country infrastructure may easily be hacked by a creative, non-stop hacking bot
- Impersonation — with the right data, you can easily mimic an individual’s way of communicating and pretend to be them online
What does GPT-3 mean for the future?
GPT-3 can create content to fit any purpose or any narrative, in mere seconds. This ranges from writing a song from one line of input (As shown above) or writing a completely fictional story. GPT-3 can even keep a conversation going (Eg. Google Assistant, Siri, etc.). The one I am personally excited for the ability of GPT-3 to help improve the mental health of people. There are already several applications of this today — such as Wysa (an AI app that talks to you about your feelings)
Conclusion:
In the end, I remember this quote from Spiderman :
“With great power comes great responsibility”
The possibility of GPT-3 is endless and as long as it is used responsibly, it can be used to improve the world we live in.
