What Is ChatGPT?
advertisement
VOICE OVER: Peter DeGiglio
The AI revolution is here. For this video, we're doing a deep dive into the chatbot taking the world by storm. ChatGPT a state-of-the-art language model developed by OpenAI that allows users to generate human-like text responses and perform various language tasks. In fact, it wrote that sentence for us. It's part of a wave of astonishing new AIs that can process natural language, like the AI art generators DALL-E, Midjourney, and Stable Diffusion. While it has enormous potential for research and efficiency, it also raises serious ethical concerns. What's your reaction to A.I.? Excitement? Fear? Indifference? Tell us in the comments.
What Is ChatGPT?
Welcome to Watchmojo, and today we’re answering the question: What is ChatGPT?
For this video, we’re doing a deep dive into the chatbot taking the world by storm.
What’s your reaction to A.I.? Excitement? Fear? Indifference? Tell us in the comments.
ChatGPT is a conversational chatbot that, in its own words, “uses deep learning algorithms and a massive amount of training data to generate human-like responses to text-based prompts”. Launched in November 2022, it was developed by San Francisco-based research laboratory OpenAI.
Its release sent shockwaves through the tech world and beyond, as users reeled from how well ChatGPT could understand and generate conversational responses. It’s also incredibly versatile: you can ask it research questions, revise text, or have it write emails, job applications, reports, and essays. It can create and debug code. And it can write fiction, including poetry. This can even be done in the style of a specific author. The results can be mixed. But ChatGPT remembers prompts and responses within the same conversation, allowing you to refine them.
The technology does have limits. Its data doesn’t go past 2021. Responses can be repetitive and lack human warmth. And they can sometimes be ‘hallucinations’ - confident responses based on deceptive data.
Nonetheless, for many, ChatGPT has been a revelation - showcasing a level of artificial intelligence that seemed far off in the future. After all, we’re still arguing with Siri and Alexa about the most basic commands … how did this happen?! Microsoft is so impressed that in January 2023, it invested a reported $10 billion in OpenAI.
However, ChatGPT’s abilities have also stoked worries about where the technology is heading.
Some of these concerns are bound up with OpenAI’s history.
The company was founded in 2015 as a non-profit by entrepreneurs, investors, and computer scientists, among them Sam Altman, Elon Musk, Peter Thiel, Ilya Sutskever, Reid Hoffman, and Jessica Livingston. Its stated goal was to “advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return”. While acknowledging the risks of A.I., Musk argued for active participation in its development to ensure safety and widespread access.
Since its founding however, OpenAI has undergone considerable changes. In 2018, Musk resigned from the board due to a conflict of interest with Tesla’s own A.I. projects. The following year, citing the expensive nature of research, CEO Sam Altman restructured the company, transitioning it from ‘non-profit’ to ‘capped-profit’. Within months, Microsoft invested $1 billion. By mid-2020, OpenAI had developed the third generation of its ‘Generative Pre-trained Transformer’ language models, upon which ChatGPT is based.
As of writing, ChatGPT remains free to use. However, there are plans to monetize it, with premium subscription services on their way. And Microsoft has announced plans to integrate OpenAI’s models into existing and new products.
These developments have provoked questions over whether OpenAI can remain true to its mission statement.
ChatGPT is part of a wave of new A.I.s that have left many people stunned. Artificial intelligence has been part of our daily lives for some time now - with algorithms guiding our search results, recommendations, and social media content. However, 2022 saw startling new powers put into our hands.
Foremost among these were text-to-image models, such as Midjourney, Stable Diffusion, and OpenAI’s own DALL-E and DALL-E 2. Just write in a prompt, and voila, you have images that are difficult to distinguish from human-created art and photographs. Again, these can even be in the style of particular artists - something that’s already prompted a class-action lawsuit.
Add in OpenAI’s multilingual speech recognition software Whisper, and A.I. seems well on its way to mastering natural language across a variety of media.
And that really just scratches the surface! Also in the works are A.I.s that can generate video from text, such as Google’s Imagen and Meta’s Make-a-Video, and others that can produce speech or music from text or soundbites. Microsoft’s VALL-E can simulate anyone’s voice from short samples.
Other chatbots are on the way too. The launch of ChatGPT reportedly caused internal panic at Google, which is working on its own answer to ChatGPT, ‘Apprentice Bard’.
A.I. is set to transform several industries in coming years. It has the power to accelerate research and take over mundane tasks.And progress could be exponential. But with great power …. Reactions have ranged from optimism to unease. In December 2022, Elon Musk Tweeted that “ChatGPT is scary good. We are not far from dangerously strong AI".
ChatGPT does have some safeguards. It will refuse to generate “content that is violent, harmful, abusive, or hateful.” Somewhat ironically, the job of teaching it to recognize such content was reportedly outsourced to workers in Kenya who were paid under $2 an hour. The experience has been described in traumatic terms. Despite this training, users have been able to jailbreak ChatGPT to violate its content policy.
In academia, a growing concern is the ability of ChatGPT to write essays, pass exams, and generate scholarly papers. While there are emerging apps to detect A.I.-written text, none are foolproof.
Then there’s the potential for misinformation - something that’s also a concern with other A.I.s, such as deepfakes, and text-to-image, -video, or -speech models. Imagine a fabricated video of the President declaring that he stole the election. This has the potential to spill over into real life violence.
At stake too are human jobs. In 2020, the World Economic Forum estimated that by 2025, machines would displace 85 million jobs, while creating 97 million new ones. A.I. will require human operators and oversight. But for many of us, it does raise the question: are we looking at a tool that will make our lives easier? Or are we looking at our replacement?
Welcome to Watchmojo, and today we’re answering the question: What is ChatGPT?
For this video, we’re doing a deep dive into the chatbot taking the world by storm.
What’s your reaction to A.I.? Excitement? Fear? Indifference? Tell us in the comments.
In a Nutshell
ChatGPT is a conversational chatbot that, in its own words, “uses deep learning algorithms and a massive amount of training data to generate human-like responses to text-based prompts”. Launched in November 2022, it was developed by San Francisco-based research laboratory OpenAI.
Its release sent shockwaves through the tech world and beyond, as users reeled from how well ChatGPT could understand and generate conversational responses. It’s also incredibly versatile: you can ask it research questions, revise text, or have it write emails, job applications, reports, and essays. It can create and debug code. And it can write fiction, including poetry. This can even be done in the style of a specific author. The results can be mixed. But ChatGPT remembers prompts and responses within the same conversation, allowing you to refine them.
The technology does have limits. Its data doesn’t go past 2021. Responses can be repetitive and lack human warmth. And they can sometimes be ‘hallucinations’ - confident responses based on deceptive data.
Nonetheless, for many, ChatGPT has been a revelation - showcasing a level of artificial intelligence that seemed far off in the future. After all, we’re still arguing with Siri and Alexa about the most basic commands … how did this happen?! Microsoft is so impressed that in January 2023, it invested a reported $10 billion in OpenAI.
However, ChatGPT’s abilities have also stoked worries about where the technology is heading.
OpenAI
Some of these concerns are bound up with OpenAI’s history.
The company was founded in 2015 as a non-profit by entrepreneurs, investors, and computer scientists, among them Sam Altman, Elon Musk, Peter Thiel, Ilya Sutskever, Reid Hoffman, and Jessica Livingston. Its stated goal was to “advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return”. While acknowledging the risks of A.I., Musk argued for active participation in its development to ensure safety and widespread access.
Since its founding however, OpenAI has undergone considerable changes. In 2018, Musk resigned from the board due to a conflict of interest with Tesla’s own A.I. projects. The following year, citing the expensive nature of research, CEO Sam Altman restructured the company, transitioning it from ‘non-profit’ to ‘capped-profit’. Within months, Microsoft invested $1 billion. By mid-2020, OpenAI had developed the third generation of its ‘Generative Pre-trained Transformer’ language models, upon which ChatGPT is based.
As of writing, ChatGPT remains free to use. However, there are plans to monetize it, with premium subscription services on their way. And Microsoft has announced plans to integrate OpenAI’s models into existing and new products.
These developments have provoked questions over whether OpenAI can remain true to its mission statement.
A.I. Everywhere
ChatGPT is part of a wave of new A.I.s that have left many people stunned. Artificial intelligence has been part of our daily lives for some time now - with algorithms guiding our search results, recommendations, and social media content. However, 2022 saw startling new powers put into our hands.
Foremost among these were text-to-image models, such as Midjourney, Stable Diffusion, and OpenAI’s own DALL-E and DALL-E 2. Just write in a prompt, and voila, you have images that are difficult to distinguish from human-created art and photographs. Again, these can even be in the style of particular artists - something that’s already prompted a class-action lawsuit.
Add in OpenAI’s multilingual speech recognition software Whisper, and A.I. seems well on its way to mastering natural language across a variety of media.
And that really just scratches the surface! Also in the works are A.I.s that can generate video from text, such as Google’s Imagen and Meta’s Make-a-Video, and others that can produce speech or music from text or soundbites. Microsoft’s VALL-E can simulate anyone’s voice from short samples.
Other chatbots are on the way too. The launch of ChatGPT reportedly caused internal panic at Google, which is working on its own answer to ChatGPT, ‘Apprentice Bard’.
Welcome to the Future
A.I. is set to transform several industries in coming years. It has the power to accelerate research and take over mundane tasks.And progress could be exponential. But with great power …. Reactions have ranged from optimism to unease. In December 2022, Elon Musk Tweeted that “ChatGPT is scary good. We are not far from dangerously strong AI".
ChatGPT does have some safeguards. It will refuse to generate “content that is violent, harmful, abusive, or hateful.” Somewhat ironically, the job of teaching it to recognize such content was reportedly outsourced to workers in Kenya who were paid under $2 an hour. The experience has been described in traumatic terms. Despite this training, users have been able to jailbreak ChatGPT to violate its content policy.
In academia, a growing concern is the ability of ChatGPT to write essays, pass exams, and generate scholarly papers. While there are emerging apps to detect A.I.-written text, none are foolproof.
Then there’s the potential for misinformation - something that’s also a concern with other A.I.s, such as deepfakes, and text-to-image, -video, or -speech models. Imagine a fabricated video of the President declaring that he stole the election. This has the potential to spill over into real life violence.
At stake too are human jobs. In 2020, the World Economic Forum estimated that by 2025, machines would displace 85 million jobs, while creating 97 million new ones. A.I. will require human operators and oversight. But for many of us, it does raise the question: are we looking at a tool that will make our lives easier? Or are we looking at our replacement?
