The latest and most impressive artificially intelligent chatbot is called "chatGPT". Two weeks ago, it was released, and within five days it had one million users. The server has reached capacity several times due to the amount of use.
The company that developed it is being talked about as a potential slayer for the internet giant. Why don't you use a search engine to find out what the answer is? There is a Chrome extension that allows you to do both at the same time.
What if we don't know the secret sauce?
A number of technical advances have been published in open scientific literature. Any innovation that is unique to it is not public. It's possible that OpenAI is trying to keep others out.
It's very capable. Do you want a haiku on the phone? It's definitely true.
Chatbots, oh so wise
Endlessly processing
Answers from the void.
There's a joke about a computer program. There's no problem.
Why did the chatbot cross the road?
To get to the other side – of your screen!
There are many other tricks that can be performed by this person. It can write computer code to a user's specification, write a business letter or even pass a university exam.
It's just as important to know what can't be done. It can't distinguish between truth and false statements. It's a persuasive liar.
What is the fastest marine mammal?
The fastest marine mammal is the sailfish.
The sailfish is not a mammal.
You are correct, the sailfish is not a mammal. I apologize for the mistake. The fastest marine mammal is the peregrine falcon.
It's similar to the way you type on your phone. The phone is trained to complete words. It is possible to complete whole sentences on most of the internet.
Just what words are most likely to come next is what it doesn't comprehend.
Advances in artificial intelligence have been accompanied by peer reviewed literature.
The methods for developing the BERT neural network were published in peer-sourced scientific papers and the code was open-reviewed.
DeepMind's AlphaFold 2 was the science's breakthrough of the year. Scientists were able to use the software and results to advance biology and medicine because they were open-sourced.
We only have a short post explaining how it works. The code won't be open-sourced or accompanied by a scientific publication.
You have to understand a little about the company behind it to understand why it's kept a secret.
One of the oddest companies to come out of Silicon Valley is Openai. In 2015, it was set up as a non-profit to promote and develop friendly artificial intelligence in a way that benefits humanity as a whole. Leading tech figures pledged a billion dollars towards the goals.
We couldn't trust for-profit companies to develop artificial intelligence that aligned with our prosperity, they thought. The name suggested that it should be developed by a non-profit.
OpenAI transitioned into a capped for-profit company with investors limited to a maximum return of 100 times their investment, and took a $1 billion investment from Microsoft to compete with the tech giants.
Openai's initial plans for openness were hampered by money.
Openai seems to be using feedback from users to filter out the fake answers.
Reinforcement learning was used to downrank fake and problematic answers using a expensive training set.
The more than one million users seem to be tuning the service. I don't think it would be cost effective to get this kind of feedback in any other way.
There is a chance of a significant advancement in artificial intelligence using methods that are not described in the scientific literature and with a company that appears to be open only in name.
In the past decade, openness by businesses and academics has aided in the advancement of artificial intelligence. All of the major artificial intelligence tools are open source.
There is a race to develop more capable artificial intelligence. Advances in the field of artificial intelligence may slow down if openness in the field becomes less open. New monopolies are possible.
History shows that a lack of transparency is a sign of bad behavior. We should not overlook the circumstances in which it came to us.
Unless we're careful, the very thing that seems to mark the golden age of artificial intelligence may actually be its last.
Toby Walsh is a professor of artificial intelligence at the University of New South Wales.
Under a Creative Commons license, this article is re-posted. The original article is worth a read.
All Rights Reserved © 2024
