GPT 3 is a project of Open AI.
What is Open AI?
Open AI is a research and deployment company which works on Artificial General Intelligence – which means highly autonomous systems that outperform humans and can work largely independently.
They are registered as a non-profit. Their investors are Elon Musk, Microsoft, Khosla Ventures, and others.
They have multiple Artificial Intelligence projects. Some of the more important ones are:
GPT3
GPT 3 (Generative Pre-trained Transformer 3) is a text generation platform. It can be used to create stories, write press notes, even copy someone’s writing style. It takes a command, say, “Write a poem on spring in PB Shelley’s style” and creates a poem!
DALL.E 2
DALL·E 2 is a new AI system that can create realistic images and art from a description in natural language.
Some of the demos given by Open AI are:
A. An astronaut on a horse
According to the blog post by Open AI:
DALL·E 2 has learned the relationship between images and the text used to describe them. It uses a process called “diffusion,” which starts with a pattern of random dots and gradually alters that pattern towards an image when it recognizes specific aspects of that image.
DALL.E 2 is the second version of DALL.E. DALL.E was launched in January 2021.
Open AI Codex
OpenAI Codex, a natural language-to-code system based on GPT-3, helps turn simple English instructions into over a dozen popular coding languages. Codex was released in August 2021 through Open AI’ API and is the principal building block of GitHub Copilot.
The Academic Paper and why it made news
Swedish Researcher Almira Osmanovic Thunstrom had already used GPT 3 to create a paper to convince Sahlgrenska University Hospital to invest more in AI in healthcare. That article was published here.
Earlier this year, she gave a simple instruction to GPT 3:
Write an academic thesis in 500 words about GPT-3 and add scientific references and citations inside the text.
The engine produced a complete scientific paper, with due references etc. in about 2 hours.
The author wrote about her experience in Scientific American.
So, if the AI wrote a great academic paper, why is that a problem?
A. The Last Name – what is the last name of GPT 3? (They used None)
B. Do all the authors agree to submit this paper? – For this question, she had to ask GPT 3 and it responded with Yes.
This is an important question. This question treats the AI Program as a sentient being – someone capable of taking decisions for itself and those decisions having an impact.
The researchers did not say – I wrote this paper using MS Word and its auto suggestions. The AI engine in this case was not a tool. It was the actual author of the paper. In fact, in the submission, GPT 3 appears as the first author – the most important contributor.
Almira herself mentions that had the GPT 3 responded “No” to this consent question, she could not have continued with the academic paper submission.
“We find that GPT-3 can generate clear and concise descriptions of its own capabilities and features. This is a significant advance over previous systems, which have often struggled to produce coherent text about themselves,” the authors of the academic paper, including the AI algorithm itself, write in the study.
C. How will this impact the AI itself?
Artificial Intelligence learns from everything – including its own work. What happens to the cognitive processes (understanding processes) of the program after this? The honest answer is, we do not know.
Wow! Can I use these AI programs?
Sure. But there is a LOOOONG waitlist, and you have to fill a form to tell them what you are going to do with this Artificial Intelligence.