Let’s understand ChatGPT and why it’s even bigger than you think
A “mind-blowing” new chatbot technology is sweeping the internet with users raving over its apparent ability to provide human-quality responses. While Some are uncomfortable with how ChatGPT uses their web content to train and learn from.
The viral AI-powered chatbot could revolutionize everything from customer service to how we search online and it may be the most important tool since modern search engines.
Let me just say up front, I see huge opportunities here. And as with all new technologies, we cannot completely anticipate the impact quite yet. To many people, the ChatGPT bot is causing panic now, but it will soon be as mundane a tool as Excel or Google Search. Automation, machine learning and artificial intelligence, these are no longer futuristic concepts, they are our reality.There will be problems and failures, but the ultimate story is creating a new chapter in human development and is transforming the way we live, work, and even think. For people, there must be a shift in mindset.
The world is changing. There’s no getting around this fact.
I believe people with grit, creativity and entrepreneurial spirit will embrace this new technology revolution, rather than stick to the status quo.
What Is ChatGPT?
In case you’ve missed the buzz, OpenAI just publicly launched its latest machine learning model, based on GPT-3.5, Chat Generative Pre-Trained Transformer, commonly called ChatGPT. The new chatbot uses deep learning techniques to generate human-like text. In other words, the robot is able to carry out natural language generation (NLG) tasks and has a remarkable ability to interact in a conversational way. The dialogue format makes it possible for ChatGPT to answer and write whatever you want quickly and prompts like an intelligent human would. It’s surprisingly useful.
What sets ChatGPT apart from a simple chatbot ?
It is the fact that it was specifically trained to understand the human intent in a question and provide helpful, truthful, and accurate answers that can appear surprisingly human.
What are potential uses of ChatGPT?
There are many potential uses for ChatGPT. They range from more direct, chatbot-type functions to much more obscure applications, and it is likely that users will explore a host of ways to utilize this technology in the future, including enhancing search engine functionality, generating coherent and well-written texts in a wide range of styles, topics and languages, automating processes like disputing traffic tickets or bills,Researching, and creating code, among many other applications.
ChatGPT has the answers, but how does it get its information?
ChatGPT is based on the GPT-3.5 language model that it has been trained on massive amounts of data about code and information from the internet, including sources like Common Crawl, webtexts, books, Wikipedia, and Reddit discussions, to help this robot learn dialogue and attain a human style of responding.
The chatbot was also trained using human feedback (a technique called Reinforcement Learning with Human Feedback) so that the AI learned what humans expected when they asked a question, and the ability to follow directions and generate responses that are satisfactory to humans. While it has no real “intelligence”( it doesn’t know what a word “means” but it knows how it is used), it can very effectively answer questions, write articles, summarize information, and more. Training the LLM this way is revolutionary because it goes beyond simply training the LLM to predict the next word.
In the demos I’ve seen over the years, the most impressive solutions I’ve seen are those which focus on a single domain. Dom, the Domino’s Robotic Unit, allows customers to place orders and customize their pizzas all within the chat, making it a cinch to buy your favorite pie. Dom has the ability to save and repeat orders and find the closest store to you. Then, there is Olivia, the AI assistant for recruiting, developed by Paradox, is smart enough to screen, interview, and hire a McDonald’s employee with amazing effectiveness.
Today, Sephora’s beauty bots boast a bounty of features. Want to browse products? Book an appointment? Try on makeup using augmented reality technology? Your virtual beauty coach can do all of that and it works very well.
This is probably one of the best chatbot examples for highlighting how a bot can take something done via mobile and make it just as good as -if not better- a real customer service assistant.
Imagine, if we built an AI that pointed to all our Marketing Strategy Research and Business Development. It would be a “virtual Sara El Khaoua Consulting” and might even be smarter than I am. (I’m thinking of starting to prototype this now)
According to Stanford University :
“GPT-3 has 175 billion parameters and was trained on 570 gigabytes of text. For comparison, its predecessor, GPT-2, was over 100 times smaller at 1.5 billion parameters.
This increase in scale drastically changes the behavior of the model — GPT-3 is able to perform tasks it was not explicitly trained on, like translating sentences from English to French, with few to no training examples.
This behavior was mostly absent in GPT-2. Furthermore, for some tasks, GPT-3 outperforms models that were explicitly trained to solve those tasks, although in other tasks it falls short.”
Large Language Models (LLMs) like ChatGPT train using multiple sources of information, including web content to really build industrial strength intelligence. OpenAI CEO Sam Altman believes these systems will “learn” from invalid data, as long as the data set gets bigger. While I understand that idea, I tend to believe the opposite. I believe that if you really want it to work reliably, you want it to source valid, deep, and expansive domain data.
But, is ChatGPT Use Of Web Content Fair?
So what I’m implying is that while AI technology has outstripped the ability of copyright laws to keep up. Some Content creators and search marketing experts are uncomfortable with how website data is used to train machines without attribution or benefit to those who published the original content used for training ChatGPT.
And that means that the consideration of whether it is ethical to train AI on web content seems to be taken for granted that Internet content can be downloaded, summarized and transformed into a product called ChatGPT.
Does that seem fair? The answer is complicated.
What are the Limitations of ChatGPT?
An important limitation of ChatGPT is that Answers Are Not Always Correct. It is specifically programmed to provide answers that feel right to humans, so the answers can trick humans that the output is correct. Many users discovered that ChatGPT can provide incorrect answers, including some that are wildly incorrect.
