The ChatGPT Generative Artificial Intelligence Revolution: Unexpected Winners
OpenAI’s ChatGPT is a global phenomenon that continues to grab headlines. The Generative Artificial Intelligence has been around since a long time. It was initially pioneered by academics, with Ian Goodfellow & Yoshua Bénigio publishing their seminal work in Generative Adversarial Networks. Then Google took up the baton and published seminal articles and patents on both GANs & generative pre-trained transformators (GPT). My first paper on generative chemical was published in 2016 and my first patent granted in 2018. The first AI-generated drugs went through phase one of clinical trials.
Forbes is one the most respected content providers in the world and most likely the most reliable when it comes to money. Forbes will not classify you a millionaire if you do not have a fortune. The site has a wealth of expert-produced longitudinal text and multimedia content, in many languages. Forbes.com has an elite team of human reporters and editors as well as a small army specializing on specific topics. It is my fifth year as a Forbes contributor, and I continue to contribute regularly in order to keep my pencil sharp. This vast human intelligence could be repurposed in part to develop internal generative tools within the Forbes empire. It could also help curate datasets, and train or benchmark third party generative tools. I would be happy to volunteer some time for such a project.
Nature and other journals within the Nature Publishing Group are considered the Olympus of academic publishing. Academics who want to publish in the elite Nature journals spend months, and even years, going through rounds of peer review and editorial. All experimental data are disclosed and all the quality of data is questioned. The thousands or millions that were spent on the experimentation is presented as a paper with supplementary materials.
Nature and other publishers can use the vast amount of high-quality data they don’t have access to to develop their own version of ChatGPT. They can also sell or license this data and restructure editorial and review processes in order to add more value to future generative systems.