Man interacting with a robotic hand in a technology lab setting.

MIT Roboticist Rodney Brooks Predicts Future of AI and Robotics

Rodney Brooks, a respected MIT roboticist, thinks people are overestimating generative AI. Brooks has been in the AI field since the 1990s and has made many accurate predictions about its future. He believes that while generative AI, like ChatGPT, is impressive, it cannot do everything a human can. Brooks warns that people often think AI can do more than it actually can. When AI systems perform certain tasks well, people assume they can do everything well. This is a mistake.

Brooks gives the example of using AI in warehouse robots. Someone suggested he use a large language model (LLM) to direct his warehouse robots. Brooks thinks this is not a good idea. He says it would slow things down. Instead, it's simpler to use data from warehouse management software to guide the robots.

Man interacting with an advanced robotic arm in a technology lab setting

Brooks also shares his thoughts on humanoid robots. He says building robots that look like humans is not always practical. His current company, Robust.AI, designs robots for specific tasks, like warehouse operations. These robots look like shopping carts with handlebars. This design makes them easier to use and fix.

Brooks believes that people are wrong to think technology always grows exponentially. He uses the iPod as an example. Early iPods doubled in storage size every few years, but eventually, this growth slowed down. He thinks this pattern applies to AI as well.

He also discusses his predictions for AI and robotics. Brooks predicted that the next big thing in AI would arrive between 2023 and 2027. He was right; large language models gained attention in 2023. Brooks also believes that AI will not make humanoid robots common for at least another 25 years. He says real progress in building robots with human-like abilities is still far away.

Brooks has other predictions, too. He thinks we won't see robots that can navigate any U.S. home before 2026. Robots that can help the elderly with multiple tasks might not arrive before 2028. And robots that can deliver packages to your door might be here by 2028.

Brooks acknowledges that LLMs could help with domestic robots in the future. But he also says this will come with challenges. He thinks the real problem is not AI's ability to do tasks but its control and optimization.

In summary, Brooks believes in the potential of AI and robotics but urges caution. He thinks people should not overestimate the current capabilities of AI systems. His experience and accurate predictions make his insights valuable.

Similar Posts