In this article from El Nuevo Dia, I shared my work at the MIT Media Lab where I was training the next generation of artificial intelligence agents to move beyond simple web searches towards more conversational intractions using common sense. As a researcher, I wanted to evolve virtual assistants so they can understand complex tasks—like guiding a user through building a chair rather than just reading a Wikipedia entry—and provide contextual advice for real-world situations like travel planning. In 2019/2020 I thought that within 10 years, these systems would be nearly indistinguishable from humans in their speech and capabilities. This article was more than 5 years ago and look where we are now with LLMs and generative AI.
