In this episode of Cyber Security Inside, Tom and Camille dive deeper into artificial intelligence with Dr. Amitai Armon, Chief Data Scientist in the IT Artificial Intelligence Group at Intel. This conversation is part 2 of a 3-part series from AI Everywhere, an internal Intel initiative and conference which includes keynotes, tech talks, tutorials, and an AI expo. The objective of this internal meeting is to encourage Intel employees to apply AI thinking and approaches to their jobs.
The conversation covers:
How artificial intelligence is being used in the industrial setting and what the goals of AI are there.
Who these data scientists are who are developing AI models, and the skills and mindset they need to have to do it successfully.
How AI is not being developed to replace humans, but to help them be more efficient and focus on what humans are better at.
What is holding up AI development and where it might go in the future.
...and more. Don’t miss it!
If you are interested in reading the article referenced and written by Amitai, you can find it at this link: https://www.calcalistech.com/ctech/articles/0,7340,L-3929057,00.html
The views and opinions expressed are those of the guests and author and do not necessarily reflect the official policy or position of Intel Corporation.
Here are some key takeaways:
AI is most often used in consumer software right now, such as in Google and Facebook. This team is focusing on using AI in an industrial setting by trying to make the machines smarter and the factories smarter. The goal is to make the manufacturing process more efficient and more useful.
The development of this AI takes people who are actually building the AI models, but also people who are building the product around the models, storing and collecting data, and engaging with customers to learn what needs need to be met.
A data scientist who builds the models needs to be passionate about data science and modeling and building products.
When talking about AI in different situations, say like a hospital, an AI is not replacing a doctor. It is just making the processing large amounts of data part easier. Humans are better at inferring from data, so they don’t need as much. But AI can process much more data. They work together.
AI does great with lots and lots of information. You can give it lots of x-rays, each with some kind of indication of whether the x-ray meant something bad or not, and the AI can learn from it. But humans can learn from 5 examples about what an x-ray should look like and understand it. AI complements people. Humans extrapolate from little information, and AI processes and interprets from a lot more information. Both are valuable.
None of the researchers and scientists on this team believe that AI will replace humans - not anytime soon. Humans still learn better and can infer. But AI can still be helpful in many ways.
There are some things an AI will never really be able to do or understand, like loving a child or feeling hungry. So it will continue to evolve to get really good at specific tasks, but it is a very long time until we have a general intelligence that can really rival that of humans.
It is important for people to know about AI and how it works, because it is already used so much in our day-to-day lives. In money, medicine, the internet, and more, AI is already used. So we should make sure it is used for good.
And who decides what is “good?” Currently, humans. Robots won’t be making those decisions.
What is preventing the development of AI from going much faster? It’s really 3 things: not having enough AI professionals (it isn’t a required course in CS degrees), computer development needing to happen, and also understanding the “secrets of nature” (Amitai). Do we know how learning and reasoning really works, or how it should work? That’s a hard question to answer.
Some interesting quotes from today’s episode:
“Not only are the machines and factories becoming smarter, the processors are also smarter. Instead of behaving the same way in every computer, they adapt themselves to the usage of the computer.” - Amitai Armon
“We need talented researchers who can do the technological breakthroughs, but can adapt them to reality - to not try to just publish a paper. I would say there are dozens of thousands of papers published in AI every year.” - Amitai Armon
“AI works differently than humans. The way that AI learns is different.. The human brain still works, learns, in a more sophisticated way than AI systems learn.” - Amitai Armon
“AI, in a sense, complements people in what it is able to do. In Intel, we believe that AI empowers people. People who use AI are able to do more and focus on what they are good at and what they like to do. Not on the tedious things that AI does better, but on the things we have advantage in.” - Amitai Armon
“The bottom line is that humans still have a learning mechanism which is far better than the learning mechanism of neural networks or other AI models. The human learning mechanism evolved over a billion years of evolution… Still, we don’t understand how humans learn, and AI learns in a much less efficient way. But it still has advantages.” - Amitai Armon
“I think it’s important for people to be educated about AI, right? It’s all around us. It’s approving our credit transactions, it decides what we see on the net, on the web. So it’s important for people to know more about it.” - Amitai Armon
“The smartest machines will probably also have no desire to conquer the world. They will just play chess or play Go… We shouldn’t be afraid of those apocalyptic scenarios of robots waking up and conquering us.” - Amitai Armon after saying that the smartest humans don’t want to rule the world, so the smartest robots shouldn’t either