Intelligently Artificial

In recent months, I have spoken about the advances in Artificial Intelligence, including a Nov 2020 Sandpiper article on GPT (Generative Pretrained Transformer, a technology that is trained on millions of documents, and generates human-like text in response to prompts), and talks both at the Del Mar Foundation and at Del Mar Community Connections.

Since its public release in Nov 2022, the ChatGPT program from OpenAI has become the fastest adopted app in history, highlighting both its immense popularity and unending intrigue. Users have gotten ChatGPT to author responses to exam questions, college-level lesson plans, admission essays, resumes, cover letters, startup pitch decks, and even novels and screenplays. Many have launched brand new businesses utilizing ChatGPT’s functionality that did not exist until four months ago. While OpenAI is upfront with their note of caution that ChatGPT-3 can produce incorrect results and that the public release is based on older documents (i.e., 2021 and earlier), the program has already scored well on SAT, and passed MBA exams, the Bar exam and even a medical board exam.

The race for technology leadership using generators and transformers is well underway. Microsoft upped its initial investment in OpenAI by 10B$, raising speculation that future Bing search and Office products will be infused with GPT. Alphabet reportedly issued an internal ‘code-red’ after the public release of ChatGPT3 and is expected to release its competitive offerings later in the year. Meta as a leader in both virtual reality and as the curator of digital conversations through Facebook and WhatsApp, can endow its bots with not only epistemologically derived knowledge but also real insights gleaned from informal or formal conversations. Chinese AI companies are considered to have an edge in computer vision and contextual awareness technologies. Venture capitalists are also betting that startups (such as OpenAI) will be significant stakeholders in this new AI race, and not just big tech.

While this level of progress and excitement is exhilarating, a common thread of questions emerge. Will this give rise to rampant plagiarism? Does it sound the death-knell for creative jobs? How to tell apart fake from real? AI language models like ChatGPT have the potential to impact the creative writing industry and lead to increased instances of plagiarism. There are steps that can be taken to mitigate these risks, such as using plagiarism detection software and developing AI models that are able to detect and flag fake content. While AI language models may change the way that writing and content creation are done, they are unlikely to completely replace human creativity and critical thinking.