A relatively recent research effort from the artificial intelligence research company “OpenAI” is the topic of much discussion among AI chatbot experts. It’s known as ChatGPT.
(The company behind the DALL-E picture-generating technology is OpenAI, a non-profit AI research lab with headquarters in San Francisco that was established in December 2015)
Conceptually, ChatGPT is similar to AI art tools because it requires little user input and creates credible synthetic media in texts rather than visuals. In reality, it may create intriguing articles, novels, and even believable and convincing poetry. You can also instruct ChatGPT to write in a particular style, just like the AI image creators.
At the moment ChatGPT was asked to provide information about Twitter in three different ways:
- In the manner of Ernest Hemingway
- In the form of Mark Twain
- In the style of a limerick
The attempts were partly successful, but the results needed to be corrected. The results were all drastically different.
ChatGPT’s natural language processing (NLP) is astonishingly effective at recognizing context. Google’s DeepMind unit used a reinforcement learning technique with human feedback (LM) to construct its dialogue language model.
Knowing the context comes from the same place as providing factual information—the knowledge that is a component of its language model.
Education, research, and other purposes are possible for ChatGPT and future AI conversation engines. But it’s also a look into how media, marketing, and corporate communication will develop in the future.