History

A brief history of AI...

OpenAI is a research organization founded in 2015 by a group of influential tech leaders including Elon Musk, Sam Altman, and Greg Brockman, among others. Its mission is to develop and promote friendly AI for the betterment of humanity, and to ensure that the benefits of AI are shared equally across society.

OpenAI has made significant contributions to the field of artificial intelligence, particularly in the areas of language processing and reinforcement learning. Its researchers have created some of the most advanced and capable language models in the world, including GPT-3, the third-generation "Generative Pre-trained Transformer" model.

ChatGPT is one of the applications of the GPT models developed by OpenAI. It is a language model designed to generate human-like responses to natural language input. ChatGPT was trained on a massive dataset of human language, allowing it to understand and respond to a wide range of topics and conversations.

ChatGPT is one of the most advanced chatbots available today, and it has been used in a variety of applications, including customer service, virtual assistants, and social media bots. Its ability to generate natural-sounding language has made it a valuable tool for businesses and individuals alike.

OpenAI is owned by its board of directors, which includes OpenAI co-founders Sam Altman, Greg Brockman, Ilya Sutskever, John Schulman, and Wojciech Zaremba, as well as representatives from Microsoft, LinkedIn, and other investors. OpenAI is a for-profit company, but it is also committed to advancing AI research in a safe and beneficial manner.

As of September 2021, Microsoft has an ownership stake in OpenAI, but the exact percentage is not publicly disclosed (rumored to be about 50%1). In 2019, Microsoft invested $11 billion in OpenAI to become its exclusive cloud provider, and as part of the deal, Microsoft also gained the right to license OpenAI's technology. However, the ownership structure of OpenAI is complex and involves multiple stakeholders, including its co-founders, employees, and other investors.

AI will not likely be free forever. It has been estimated that each ChatGPT query costs about $0.36! It is expected that the cost per query will come down drastically over the coming months and years. However, with the dramatic use of servers, processing, and data storage it is likely to be costly.

Other interesting facts...

How does ChatGPT AI Work?

“ChatGPT is based on a deep learning architecture called the Transformer Network.” The GPT stands for generative pre-trained transformer. It works similarly to how a smartphone suggests the next word in a text or proper spelling of a word. A computer algorithm is fed an extremely large amount of data. Based on this data, the computer analyzes word and sentence structure and offers up the most likely next word or sentence to follow. It does not think or reason with the data. It is simply making a very educated guess based on the provided data. 1

ChatGPT is not sentient. It simply has access to a large amount of text information and make guesses based on probibility. But instead of always picking the most likely option, it picks the next word (or character) with a bit of randomness. This leads to made up "hallucinations" or even made up words at times. Though, it has access to some amazing information and is very good at crafting blogs, essays, or other publications. 

Companies Who Are Competing in AI

The Competitive Landscape of AI Companies

OpenAI, a leading AI research organization, has developed ChatGPT, a powerful language model. Microsoft, a key investor in OpenAI, has invested over $11 billion into ChatGPT and plans to integrate it into their product line, including Bing search engine and Office suite.

Since its launch in November 2022, ChatGPT has gained significant traction, boasting over 100 million users in February 2023 and maintaining over 10 million daily active users. The dataset used for training ChatGPT contains data up to September 2021; however, other GPT models may have access to more recent information.

Anthropic is an emerging AI company that aims to develop AI models with fewer biases and improved safety compared to existing solutions like ChatGPT. By focusing on reducing risks and ensuring responsible AI deployment, Anthropic seeks to address some of the concerns surrounding AI technology.

Bard is Google's proprietary AI language model, currently not available to the general public. Preliminary reports suggest that Bard may produce more relevant and succinct results than ChatGPT, but it remains to be seen whether this performance holds up outside of lab or limited-use settings.

Microsoft's Bing search engine utilizes a customized version of OpenAI's GPT to improve search results and provide users with enhanced information retrieval capabilities.

Quillbot is an AI-powered tool designed to refine AI-generated content, making it sound more human-like in tone and structure. This technology helps evade AI detectors and improve the overall quality of AI-produced text.

The underlying technology of GPT models has proven to be relatively easy to replicate. For example, Stanford researchers have successfully developed a competing model to ChatGPT using Meta's (formerly Facebook) Open Token model and training data derived from ChatGPT's own responses. While OpenAI incorporates various safety measures in their ChatGPT product, independent developers may not include such protections, raising concerns about potential misuse of AI technology.

With the rapid development of AI language models, companies like OpenAI, Anthropic, Google, Microsoft, and Quillbot are pushing the boundaries of natural language understanding and generation. As competition heats up, concerns about AI safety, bias, and responsible deployment become even more critical to address, ensuring a future where AI serves humanity responsibly and ethically.

Things to Consider

Privacy 

ChatGPT does use your data in a few ways. If you are considering using ChatGPT, please read their privacy statement and understand how your (and your students) data is likely to be used. 


Are ChatGPT conversations private or are they made public? The privacy policy states that the data you provide to ChatGPT is used to improve ChatGPT’s AI. Information is not shared with third parties (unless required by law) and the data is not linked to individuals. Their servers use state of the art security to protect the data. However, servers like this do get hacked from time-to-time. It is good practice to not share personal information, such as your address, financial data, passwords, etc. Software developers have the ability to see the data you provide and their servers may one day be compromised. Use caution. 1

Information and Bias

In the age of AI, we will need to reconsider what is truly important for our students. AI is a technology tool that is owned and manipulated by companies. It can be used for propaganda or marketing and it can be deliberately or mistakenly full of bias or fundamentally inaccurate information. Students, now, need an extremely sophisticated ability to identify bias, identify sources, and evaluate the information to which they are being presented. 

Writing Tone

AI does a fantastic job of writing, particularly in academic settings. But its writing is, currently, somewhat algorithmic in tone and structure. This may change in the coming months, as AI becomes more and more sophisticated. 


The tone is going to change over time. There may currently be an AI echo in the system. ChatGPT was trained on “human text” such as wikipedia and some blogs. However, another older model of AI has been used for years in wikipedia and in blogs. So the AI may be learning what AI writes like and thinking it's human. 

AI Detectors 

There are many AI detectors available but they are not very accurate and can be fooled. If used, they should be used as a data point and opportunity for discussion, rather than catching students who are using AI. Remember that the deficit mindset that we encourage out students to avoid should also be avoided by us, teachers. Keep in that growth mindset and work within the constraints of AI to encourage learning. 

How they work

AI detectors “use machine learning algorithms to analyze text for features that are indicative of AI generated content.“ 1 They look for patterns and features in sentences that look like typical AI generated text. They also look for known AI generated text and compare it. 


AI is evolving very fast and there are challenges with AI Detectors. The first detectors were created by showing AI a large about of AI generated text, then showing it a large amount of human generated text, and telling it to find the patterns and differences. However, AI has been out for years and the Davinci model of AI has been generating content that may have shown up in the human sample of text. This makes the detectors less effective. Google’s AI tool is 40-60% accurate. 1

AI Being Used to Defeat AI Detectors

Quillbot is a tool that will take AI generated content and make it more human in tone and structure, to avoid AI detectors.  

Opportunities and Things ChatGPT Does Well

Some areas where ChatGPT is being utilized include:

By using ChatGPT responsibly and ethically, students can enhance their learning experience, develop valuable skills for their future careers, and explore creative applications of AI technology.

Please see my Google Sources Doc for citations and resources.