What impact will the future of AI generation have? |MIT News

When Openai introduced Chatgpt to the world in 2022, it brought the generated AI to the mainstream and started the snowball effect, allowing it to quickly integrate into the industry, scientific research, health care, and the daily lives of people using the technology.

What’s next for this powerful but imperfect tool?

With this in mind, hundreds of researchers, business leaders, educators and students gathered at MIT’s KRESGE Auditorium to attend the inaugural MIT Generation AI Impact Inpectium (MGAIC) workshop on September 17 to share insights and discuss the potential future of AI generation.

“This is a pivotal moment — generalized AI is moving fast. It is our job to make sure that, as the technology keeps advancing, our collective wisdom keeps pace,” said MIT Provost Anantha Chandrakasan to kick off this first symposium of the MGAIC, a consortium of industry leaders and MIT researchers launched in February to harness the power of generalized AI for the good of society.

MIT President Sally Kornbluth highlighted the key need for this collaborative effort, saying that as technology advances, the world is counting on teachers, researchers and business leaders like MGAIC to address the technical and ethical challenges of generating AI.

“Part of MIT’s responsibility is to keep these advances coming. …How do we manage (generate AI) magic so that we all can confidently rely on it to achieve critical applications in the real world?” Cohenbruce said.

For keynote speakers by Yann Lecun, META’s chief AI scientist, the most exciting and significant advancement in generating AI will most likely not come from the continuous improvement or expansion of large language models all over camels, GPT and Claude. Through training, these huge generative models learn patterns in huge datasets to produce new outputs.

Instead, Lucun and others are developing the development of “world models” that learn how these “world models” are learned by interacting and interacting with the world around them through sensory input.

“A 4-year-old saw as much data as the largest LLM through his vision. … The world model will become a key component of future AI systems,” he said.

Robots with this world model can learn to complete a new task themselves without training. Lecun believes that the world model is the best way for companies to make robots smart enough to be useful in the real world.

But even if future generative AI systems do become smarter and more humane by combining world models, Lecun doesn’t worry about robots escaping human control.

He said scientists and engineers will need to design guardrails to maintain future AI systems, but as a society, rules that we have designed rules to align human behavior with the common good have been doing so for this millennium.

“We will have to design these guardrails, but by construction, the system will not be able to escape them,” Lecun said.

Keynote speakers for Amazon Robotics’ chief technologist Tye Brady also discussed how generative AI can impact the future of robotics.

For example, Amazon has incorporated generated AI technology into many warehouses to optimize how robots drive and move materials to simplify order processing.

He hopes that many future innovations will focus on the use of generative AI in machines that make humans more efficient.

“Genai is probably the most influential technology I’ve witnessed throughout my entire robotics career,” he said.

Other presenters and panelists discussed the impact of generative AI on businesses, from large businesses such as Coca-Cola and simulation devices to startups such as Health Care AI Company Abridge.

Some MIT faculty members also talked about their latest research projects, including the use of AI to reduce noise in ecological image data, designing new AI systems to mitigate bias and hallucinations, and enabling LLMS to learn more about the visual world.

After exploring new generative AI technologies and discussing their impact on the future, Professor Patrick J. McGovern, co-led by MGAIC faculty, said he hopes attendees can make “different and keep the possibilities urgent to make that possibility real.”