We believe that it is important to deploy solutions even if the underlying AI solutions are not one hundred percent robust. This can be accomplished by employing scaffolding from more traditional techniques (such as feature engineering and bespoke scripting) as well as finding applications in which partial solutions still add significant value and where incremental improvement can be part of the overall experience. Systems where the AI is seen to be incrementally learning and where user teaching is encouraged fall into this category.
Our goal is to develop transformative experiences that utilize AI to engage audiences with content in ways that would not have been possible by other means, especially at scale. These experiences lie somewhere on the continuum between traditional linear content and interactive video games, while also offering unprecedented opportunities for customization and personalization usually available only in sophisticated and difficult to learn content creation tools. An example would be an interactive cartoon series that plays out like an improvisational theatre production. Audience suggestions would be seamlessly and coherently integrated into the action, offering a unique and ongoing storyline for every audience in a way that is customized to their whims. These experiences could be highly interactive or invisibly personalized. They promise to be deeply social and individually resonant.