MIT’s Generative AI Week began with a symposium on November 28, titled “Generative AI: Shaping the Future”. The keynote speaker was Rodney Brooks, co-founder of iRobot and former director of the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT.
During his address, Brooks cautioned against overestimating the capabilities of generative AI, which forms the basis for power tools like OpenAI’s ChatGPT and Google’s Bard. He stated that no one technology has ever surpassed others entirely and highlighted the potential dangers of hype leading to failure.
Generative AI uses machine learning to create new content which resembles the data it has been trained on. It exhibits impressive abilities such as creating human-like writing, translating languages, crafting realistic images from text prompts, and coding.
MIT President Sally Kornbluth, during her opening remarks, pointed out how MIT faculty and students have been utilizing generative AI for community improvement. Examples include the Axim Collaborative, an educational initiative by MIT and Harvard using generative AI to help underprivileged students. MIT also plans to award seed grants for 27 projects concerning the societal impact of AI.
She emphasized the importance of collective action from academia, policymakers, and industry toward safe integration of advanced technologies like generative AI. Kornbluth asserts that this integration aligns with MIT’s mission and could help solve numerous global issues.
Daniela Rus, Director of CSAIL, stressed the blurring boundaries between science fiction and reality due to the emergence of generative AI. She mentioned the need to enhance businesses and ensure sustainability using AI tools, instead of creating new content.
The symposium also discussed concerns with generative AI, like teams abandoning breakthrough projects in favor of AI advancements, venture capital firms chasing high profit margins, or an engineer generation forgetting other software types. Brooks underlines how both factions – those who see AI as a problem solver and those who believe it causes more problems – tend to overrate the technology.
A panel discussion including MIT faculty followed Brooks’ talk. They discussed potential research directions around generative AI, such as integrating perceptual systems. Panelists also emphasized the need for policymakers and public engagement to ensure responsible AI deployment.
A roundtable discussion later included discussions on developing AI models that could surpass human abilities, with a focus on trust for safe real-world applications.
