At the end of November, OpenAI released an artificial intelligence (AI) chatbot tool called ChatGPT, which has been called an “iPhone moment in AI” due to its potential to disrupt the technology industry. Although chatbots are a relatively common tool in today’s world, ChatGPT differentiates itself through its ability to instantaneously explain complex concepts in simple terms and to generate creative content from scratch, among other things.

According to The New York Times, “ChatGPT is, quite simply, the best artificial intelligence chatbot ever released to the general public.” The tool is currently in a research preview and is free to use.

Based on a generative pretrained transformer (GPT) language model, ChatGPT is trained to generate human-like text through the use of machine learning algorithms that analyze an enormous amount of data and learn the patterns and structures of the language. The model is designed to interact in a dialogue format that, according to OpenAI, “makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”

The uses of ChatGPT are wide-ranging and flexible, from explaining scientific concepts, coding software programs, providing gift ideas, telling jokes, and writing original content, such as poems about a beloved pet (in almost any style you can imagine) and answers to college essay questions (i.e., cheating in school!). It will even remember and learn from your past “conversations.”

We decided to test ChatGPT’s blog writing abilities and gave it the following prompt: Identify in five bullet points the top legal issues with using generative pretrained transformer language models.

ChatGPT responded as follows:

Here are five potential legal issues that may arise when using Generative Pre-trained Transformer (GPT) language models or other similar AI tools:

1. Copyright infringement: If a language model generates text or content that is based on existing works, it is possible that the generated output could be considered a copy of the original work. This could result in a claim of copyright infringement if the generated content is published or distributed without the permission of the copyright owner.

2. Trademark infringement: If a language model generates content that includes trademarks or brand names, it is possible that the use of these marks could be considered infringing if it is done without the permission of the trademark owner.

3. Defamation: If a language model generates content that is false or misleading, and this content is published or distributed, it is possible that the person or entity responsible for the content could be sued for defamation.

4. Privacy and data protection: If a language model is used to collect or process personal information, it is important to ensure that this data is handled in accordance with relevant privacy laws and regulations.

5. Ethical implications: AI systems can potentially perpetuate or amplify existing biases, and it is important to consider the potential impacts of their use on society and individuals. It is important to consider the ethical implications of any technology or tool that you use, and to take steps to minimize the potential negative impacts of their use.

Not too bad, ChatGPT! Although, we should note that OpenAI readily admits that “ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers.”

It is not hard to imagine how game changing this technology could be, especially when the next iteration of the AI model (GPT-4) is released. For now, it is fun to explore the many creative, smart, and weird uses of the tool, but it is also important to start thinking of the potential impacts and associated legal risks that these AI tools will have on businesses and individuals.

Happy New Year!