Last weekend I had the great opportunity to pitch my startup, Fireflies AI, at the MIT Tech Conference. This year’s topic was “The Rise of Artificial Intelligence.” In attendance, were some of the brightest and greatest minds in AI. From Ray Kurzweil to Rob High the day was filled with great speakers and panels on what the future would look like. Leading experts and professors in the field held exciting discussions on what the world might look like in 2025 and how to make sure AI doesn’t harm us.
Why is AI hot again
In the 1950s, AI pioneer Marvin Minsky had made great breakthroughs in the field estimating that by the 1960s we would have a “complete AI.” It was not until recently that these thoughts have resurfaced. Recent advancements in the field have generated a lot of hype. From IBM Watson winning Jeopardy to what once thought was impossible, an AI beating the game Go. Just this year, a team at Google was finally able to create an AI to concur the game Go. Developments like these are what have started stirring the great debate of AI once again.
Rob High, the CTO of IBM Watson kicked off the event and gave a dazzling overview of how far Watson has progressed, from winning Jeopardy a few years ago, to today where its trying to help doctors cure cancer. High’s vision of the future is unlike ours. Rather than have an AI replace humans, he sees AI assisting humans evolve. He wants AI to help make the world a better place. In the classical sense, AI is software that recreates human capabilities. However, Watson is trying its best to amplify human cognition. The next big challenge for IBM’s Watson is trying to understand emotion and expression.
The future holds a lot of surprises
The first panel focused on what the world would look in 2025. Imagine a world where ten people are collaborating together, sharing your information, and trying their best to help you. Now replace each of them with an AI. This fantasy could soon be reality. Over the past few months, and heading into the decade you will see a lot of AI agents who are really good at certain verticals. Some will be great at socializing while others will make sure your meetings are organized. These agents will start talking to each other using NLP (Natural Language Processing), not even differentiating to each other that they are an AI, but acting like humans. Conversations that sound like humans talking instead will be AI agents communicating.
On the topic of AI agents, one of the panels during the conference agreed that they only way for AI agents to achieve quick adoption and growth would be in the industries where they can quickly monetize. While this is a common trend in new technologies, it will be interesting seeing how everyone applies AI in the next few years. AI has already had a large impact in enterprise applications. For consumer applications will we see more than the “complete AI” attempts like Cortana, Google Now, or Siri? Or will we see a new approach to how AI has social impact?
We will see AI surpass humans in 2029
Ray KurzIn the hour Kurzweil spoke, he started with the memory of his friend and AI pioneer, the late Marvin Minsky. Minsky paved the way for AI and without him the world would have been a different place, one with less imagination. Kurzweil rewinded back to the mid 20th century, spending most of the time talking about how technology trends follow exponential curves and not the linear ones we all think they do. He stated he had been able to predict numerous technologies just by using this exponential curve. He went into talking about how advanced the biotech industry was and that we were very close (5–10 years) from seeing reproducible organs and immune system “upgrades.” Diving into the connected mind, Kurzweil explained how in the next few decades we could have our brains connected to artificial networks to help us process information faster. Assisting our brain’s slow biochemical process by introducing silicon. To conclude his talk about how fast technology moves, he ended the presentation with his estimate on when complete AI will finally become reality, in 10 years.
If you imagine where you were ten years ago, you might have been lucky to have a candy bar cell phone. The difference between 2006 and 2016 are drastic. By Kurzweil’s estimates, our lifestyle in 2026 will finally see the introduction of a complete AI. You might be thinking that we have been talking about complete AI since the 1950’s and it still hasn’t happened. A survey done at a recent AI conference by an MIT professor showed that majority of AI experts see “complete AI” finally becoming reality in the next 10–20 years, with a few saying in the next 5 years and others saying it will never happen. Our future might be one of surprises and unexpected outcomes. One that could lead to the “doom of humanity.”
AI will have its own Treaty of Paris
Terminator. Skynet. These are often the images that come to our head when we think of the future and AI. We fear the harm it could cause society. The last panel, consisting of leading AI professors focused on making sure AI stayed safe, also known as “safe AI.” They are already working on an AI treaty to make sure AI would never be used to harm. The treaty is similar to the treaties signed against biochemical warfare and nuclear weapons. In order to ensure the future, they are working with the world’s leaders to guarantee that AI is used to help humanity.
Another counter argument to the topic of harmful AI is that humanity today is not safe, AI will just amplify humans since humans are the ones making AI. Another interesting idea brought by University of Wisconsin professor, Bill Hibbert, is that all AI should remain open so experts all over the world could analyze it to ensure its not harmful.
I would like to end with that.
Over the past century, AI research has remained open. Allowing companies like Google and Facebook to feed their custom data sets into the same algorithms that you yourself could use on your computer today. We have seen great applications of AI from analyzing images to to even recreating images in paintings with the style of your favorite artist. Over the next few decades we will start seeing breakthroughs in technology not thought of before. With it will come a lot of social unrest and inequality. As long as we work together, for each other, we might see the fantasy of 20 hour work weeks and an intelligence simulated society shooting for the stars come true.
