April 02, 2020 at 03:00PM
[voices_in_ai_byline]
About this Episode
On Episode 110 of Voices in AI, Byron speaks with Didem Un Ates, the Senior Director of AI customer and partner engagement for Microsoft about Artificial Intelligence.
Listen to this episode or read the full transcript at www.VoicesinAI.com
Transcript Excerpt
Byron Reese: This is Voices in AI brought to you by GigaOm. I’m Byron Reese. Today, my guest is Didem Un Ates. She is with Microsoft and her title is the Senior Director of AI Customer and Partner Engagement. She’s been there for several years. She holds two degrees, including in electrical engineering, from the University of Pennsylvania and she has an MBA from Columbia as well. She’s joining us from London. Welcome, to the show, Didem!
Didem Un Ates: Hi, Byron. Thanks for having me.
I always like to start with the same group of questions, which begins with: What is artificial intelligence and why exactly is it artificial? What’s artificial about it and what is intelligence for that matter?
Thank you. The way I try to explain it to my customers, partners and other individuals like students at schools – universities or high schools, is basically: artificial intelligence is a way of mimicking our brain. Intelligence makes sense of things around us. It’s how we process our environment, how we make sense of it, make these connections between the past and the future and the present, that’s called general intelligence. Then we also have specific intelligences, which is all very specific functions like object recognition or speech recognition. ‘Artificial’ is trying to mimic this with technology, with algorithms.
Well, it’s interesting you’re saying the word ‘mimic.’ Is that to imply it’s not actual intelligence? It’s just doing something that can emulate intelligence or do you actually think it’s smart?
No, it’s definitely smart and it’s – in some cases, the specific intelligence that I referred to, some call it weak AI, is actually already smarter than humans in those areas. Microsoft actually was the first to surpass human intelligence in speech recognition, translation, object recognition etc. Yes, some of these functional areas are already very smart and even smarter than humans, but the general AI, or the strong AI as some like to call it, is around a five year old’s intelligence level. That’s why I call it ‘mimic.’
When you say – you think we’re at a five year old [level] for general intelligence, is that really the case? It seems to me that we have this one trick that’s been working pretty well for a while, which is machine-learning, where we take a bunch of data about the past and we study it and we make projections into the future. That seems to be a really – not a very generalized tool. There are a lot of things where the future’s not like the past. The word ‘banana’ is said the same way tomorrow and yesterday so it’s a really good thing that you could do that.
Things like creativity and other sorts of things we associate with general intelligence, are they even solvable that way? When you say we’re at a five year old, that means maybe next year we’ll be at a six year old, at a seven, and then in 15 or 20 years, we’ll be at a teenager. Is there a limit to our one little trick we know here and what it’s going to be able to do?
These are great questions, and I think similar to you, Byron, I’m obsessively reading about AI and trying to get different perspectives on the experts, mostly at the universities but also the industry. To me, when I say – or when we read about the general intelligence is around the age of a five year old human being right now, all it means is as we improve the algorithms around AI and ML, we mimic the human learning – human brain and it is at the level of a five year old human being. Some of course predict actually that general AI are actually racing to reach an adult human mind. It’s my personal view, not Microsoft’s view by far, but my personal view is: yes, AI/ML will reach adult intelligence, but when this will happen is a big question.
To your point about teenager years, some predict that it will be happening in five years; others are saying it won’t happen in a century. The average at least in my reading and research is somewhere around 15 to 25 years. This is completely my own – let’s say doing my own homework. This is quite serious because it has many implications in terms of let’s say, automation or impact on society, jobs, changes, exciting things coming, and also lots of integrations that we should proactively manage in terms of responsible ethical AI, which we are, as Microsoft, very, very, serious about.
Listen to this episode or read the full transcript at www.VoicesinAI.com
[voices_in_ai_link_back]
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.
No comments:
Post a Comment