The Origins and Advancements of this Amazing Piece of Artificial Intelligent Technology
Creating a Machine that acts and behaves like a human was once thought to be a top aspiration within the field of Artificial intelligence. Well with the power of machine learning and the advancements in modern technologies we are closer now to a human like chat bot more than ever before.
Once the Chatbot has been perfected all that needs to be done is placed in a humanoid frame. Then we will have androids (scary stuff right?). Let’s not get ahead of ourselves just yet.
Not so long ago Chatbots were regarded as exclusive applications that were only viable to the worlds largest tech companies, this was due to their costliness and the sheer power they require to run.
Well today we have the Introduction of the Large Language Models (LLM) in some cases you can run a LLM on your home computer, well if it’s powerful enough of course. It’s really just a matter of time before these LLM models will be made more portable so anyone can use them. Heck we already have API’s that we can connect to remotely, to use them. If you are old enough to remember iRobot (2004) this was an example of remote access for the robots seen on this movie.
With Chatbots being everywhere in the news today we would forgive you if you thought it was a new piece of technology. Here’s the shocker, in fact, chatbots have been around since the early days of the computer! (1966) So people have been working on this form of AI for quite sometime.
In this article we delve into the origins of the humble Chatbot and what advanced in technology have happened over the years.
The Early Origins of the Humble AI Chatbots
The first Chatbot created was nicknamed – Eliza – Eliza is a remarkable 55 years old. So as you can see AI chatbots have been a long time coming.
The Creator of Eliza is Joseph Weizenbaum he was a German Computer Scientist. Back in 1966 the programming methods were a long way off, of what they are today. Even though Eliza still was capable to prove that computers can try to mimic human text conversations. Of course Eliza relied heavily on processing input statements through an set algorithm created by Weizenbaum these were known as scripts. This give people the impression that Eliza was able to understand text. But as these were limited it was only a matter of time before the chatbot showed it’s limitations in the conversations.
To simplify the process we can sum up this chatbots behaviour with the following example. Eliza would have had a data bank of responses that respond when keywords are given.
If “hi” responds with “hi”
If “what is your name” respond with “Elizia”
Now the issue would arise if say you spelt a word wrong or asked Eliza a question that was not stored in her databanks. This issue today is solved with deep learning and machine learning.
After Eliza there were various creations of the Chatbot that were created up to the 80’s. One of the most famous creations was known as Parry Parry was built in 1972 By a Professor of Psychiatry at Stanford University, USA Known as Kenneth Colby.
It was Parry’s job to try to emulate a person with schizophrenia. It in theory is a chatbot quite similar to Eliza but it has a built in conversational strategy aimed at improving it’s responses. The fun thing about the creations of Chatbot’s is like people love to pit them against one another;. There has been numerous conversations between Eliza and Parry to see how they react with one another, it’s always interesting to see this interaction.
Parry managed to fool 52% of psychiatrists that analysed it’s responses. But when we consider it was written by a professor in the field, he could quite easily put responses down that would make Parry look like a real patient. Times were different back then, when this form of chat is viewed today it does of course seem quite primitive when compared to the success stories of ChatGPT and other LLM’s.
Here is a breakdown of some notable chatbots that have been created over the years:
There was many more created but these deserver proper notification for their advances in the field.
Into the Modern Day: Machine Learning and Natural Language Processing
When it comes to todays standards, Chatbots have improved by a huge standard. This is mainly tanks to Machine learning (ML_ and of course Natural Language Processing (NLP).
Without the use of ML we would not be able to train AI algorithms on large amounts of data. NLP annotation and data labelling has also made it much easier for AI to abosrb this information. This has of course lead to more accurate conversational chatbots. They come across more natural in their responses some may even say life like. But please note this is not AGI just yet, but it might be in the future.
When people think of the modern chatbot they might think up of names like Siri, Alexa and the other big names. But this is not really a chatbot, it is more of a speech recognised software. Just think of this kind of AI as in it will listen to a word, grab said word for a dataset and reply with the response in the data set. Kind of like Eliza with an upgrade to listen to voice.
AI uses a mix of machine learning, LLMs and NLPs in order to understand speech and text. Without these advances in these technologies we would not have some of the more advanced chat bots of todays age.
The true benefit of machine learning and deep learning is that it require less human interaction and in theory the chatbot can learn it’s self. Over time of course. I bet you were not aware that there are over 100,000 chatbot in existence just on Facebook alone. Crazy right?
There are some remarkable chatbots available to use online today, such as:
- ChatGPT – The world most advanced Chat Bot
- Bing Chat – Microsoft answer to Chat Bots
- Bard – Googles own LLM.
Some of these Chatbots are trying to be either an AI assistant or a companion, you can clearly see which direction some of these companies want to head it. We expect to hear a lot from these 3 of the coming months and even years, as they fight for the number one spot.
So What’s The Future of Chatbots Like
Chatbots already handle complex user responses. Offers users help, write articles and even create digital content. But there is so much more than can do for us. As chat bots are implemented into our daily lives and have access to more online services they will be able to pretty much be our very own personal assistant. They will also be able to assist us with a wide range of tasks.
We are already seen the impact AI is having on science, agriculture and even manufacturing. Maybe we might also see this implemented into the hospitality doctor and even our healthcare and education systems.
All we know is that AI Chatbots are a good addition to humanity as long as they do not take over our day to day lives and are used as an aid to improve them.
Conclusion: The Evolution of AI Chatbots
Here at So Artificial we have no doubt that Chatbots will play a large role in our lives in over the next decade. Weather this is good or bad, well only time will tell. This might be why there is a lot speculation around regulating these new systems.
Chat bots can be useful for a wide range of tasks and are already replacing some peoples jobs, which is not a good sign. Let’s hope in the near future we have a more balanced approach to chat bots and we are living in harmony with them.
What do you think of Chatbots are they a good asset to have or a plague on mankind? let us know in the comments below.