The best intel from CEOs and experts on how AI is changing business.

Insights: Artificial Intelligence

[MUSIC PLAYING] ADAM LASHINSKY: Now, AI is being incorporated into all sorts of things. The Alexa Echo, or Google Home, or Siri, that's very garden variety AI. This is really first inning AI. MARY BETH LAUGHTON: We always try to not just do technology for technology's sake, but actually solving a consumer problem pain point. We feel like we need to have the right data and the right talent to actually really take advantage of AI, so we're kind of focused on that first. But we're also experimenting in a couple different ways of getting some AI out there just to test. ANDREW NUSCA: Most CEOs are asking themselves what can I do with AI? How can I incorporate this? This is clearly the future, but they don't really know. And that's part of the problem. It's very abstract. Most of the applications you see today involve machine learning. Google image search, that's machine learning comparing things. The certain kind of ad on Facebook, that's machine learning trying to tailor those ads to the preferences that you have expressed on Facebook. The more advanced stuff, that's not really in the wild yet but for the largest tech companies. NAVEEN RAO: So artificial intelligence actually has its roots back in computing-- Alan Turing even. He was really trying to recapitulate pieces of intelligence. What made a machine intelligent? And he recognized very early on that, that lived in the information space. Processing information to make it useful somehow. The term artificial intelligence was coined in the '60s. We started to understand the brain was definitely electrical. In the '40s, we actually defined how it works electrically, how information is represented, what we call spiking. That actually was concurrent with a lot of innovations on the computing side. Because we said, well, brains are information tools, computers are information tools, and there's always this parallel. The first computer science department and the first neuroscience department, started a year apart. And we figured out that neurons represent information a certain way. We tried to simulate those things in synthetic systems. We thought it would be really easy because we figured out all these principles from the brain. It turns out it's a little more difficult than that. So the field kind of had its ups and downs until computing got to a point where it was dense enough. And the data set sizes got to the point where they're big enough. 2010, 2011 time frame is when neural networks met Moore's law met big data. This whole field exploded. It's going to change with time, but overall the goal of a learning system is to find useful structure and data. ADAM LASHINSKY: Prominent AI researchers have said that AI will be like electricity. Electricity was a big deal. It was this new invention that people talked about. Nobody talks about electricity anymore and that's where we are with AI. X years down the road, sooner rather than later, AI will be infused in all these things in ways that we couldn't possibly fathom today. ANN CAIRNS: The way that you use a telephone is different from the way that I use a telephone, the way that we actually key. We have artificial intelligence that can detect your behavior. So if I left my phone there and you picked it up and tried to make a payment, it would know it's not me. JEREMY KING: When we're unloading a truck full of strawberries, one of the drivers can open up the case and take a photo. We can determine whether it was sent sustainably and how long that fruit is going to last. MARY BETH LAUGHTON: I think AI can be the future of personalization. And I think personalization is where retail and beauty are going. We have what's called Sephora Virtual Artist, which is our augmented reality beauty experience. We're using computer vision, so it can map your facial features and know where to place the makeup when you look at it virtually. SPEAKER 1: We've actually been using AI to really reshape the entire end-to-end shopping journey for the customer. Increasingly, we use AI to show up through targeted display ads in places where the consumer is shopping. So if you let us know that you're interested in certain lifestyles and certain items, we'll make sure that we personalize the recommendations that we make for you. CHRISTOPHER HEISER: Our customers use a lot of AI. The self-driving systems use it for object detection. We work with people that are perceiving the occupants. Are they happy? Are they scared? Are losing trust with the vehicle? There are already dozens of applications of AI. I only expect that to increase. ANDREW NUSCA: There is no technological development with more potential than artificial intelligence. Because it is a suite of capabilities, it's going to solve simple things in our lives and the extraordinary things. What is really exciting about that for the business community, there's money to be made. MICHAL LEV-RAM: There's so much excitement around AI, rightfully so. But if we're not thoughtful about how we go about building these systems, the outcome could be disastrous. ANN CAIRNS: If there aren't enough of us engaged, and I don't actually even just mean women, I mean a diverse group of people, then we're going to get computers who think like some bro culture on the West Coast. MICHAL LEV-RAM: Bias can be inherent to AI and to algorithms. If we allow it to be, comes down to people. If those teams aren't reflective of humanity, then the algorithms aren't going to be reflective of humanity and that has hugely negative ramifications. ANDREW NUSCA: One of the biggest concerns is that workplaces will lose their humanity if we keep incorporating more and more AI into the office. CHRISTINA CULCLOUGH: There's a lot of algorithms that are being put into place that help the companies in their recruitment processes, who are they going to hire, who they're going to fire, and so on. But the big question is what data are they using? Are they bringing in your social media data, your private data? Do they get access to your whole file, your financial data? CHRISTOPHER HEISER: The amount of information that we freely give or unknowingly give to corporations and governments is massive. AI will allow those groups to manage in a very efficient way. Whether that's something that's good for me or bad for me, I think, depends on who's doing it and how. CHRISTINA CULCLOUGH: We have all sleepwalked into this current situation. We haven't really understood what it was we were giving away when we were using all of these services, when we're using social media. What is it we are paying with to get all of these services? CHRISTOPHER HEISER: We've got people arguing that AI is great and other people arguing that it's the end of humanity. The answer is always somewhere in between. It is something we have to think pretty deeply about. CHRISTINA CULCLOUGH: Every single technological development has changed jobs. I mean, that's nothing new. Where I think this whole transformation is new is its magnitude. It's going to affect all of us. ANDREW NG: I am concerned that these hype discussion about evil AI, killer robots, it whitewashes the real challenges. And I think job displacement is a big one. CHRISTINA CULCLOUGH: A lot of jobs will change, meaning that we will have to learn new things. And this brings us to the whole question of re and upskilling. This is going to be one of the major challenges of our time, making sure that everybody can follow the developments and have an opportunity to thrive in the labor market. JEREMY KING: What happens is jobs shift? We have about 18,000 new jobs just for grocery picking and delivery services. JUSTIN TRUDEAU: Robotics, AI, technological advances are rapidly disrupting and transforming the nature of work and the workplace. We know what happens with industrial revolutions. There is a transition time in which a whole bunch of people get disrupted right out of their jobs. Once it's through that transition time, there are better quality jobs and more jobs for everyone. CHRISTINA CULCLOUGH: Predictions are that AI will create many more jobs than it will destroy. If those jobs that will be created are bad jobs, are paid under their living wage, are precarious, then that is nothing we should celebrate. JEREMY KING: I don't know if you've seen this picture of sheep problem. If you showed a computer scientist, three or four years ago, a picture of a flock of sheep and ask them to build an algorithm to count how many sheep were in the picture, they would have told you it's impossible. Now, that's trivial. Overnight that changed. When I think about how fast AI, machine learning is going to move and that's the kind of speed we're looking at, that's exciting. NAVEEN RAO: Right now the depth of expertise required is extremely high. CHRISTOPHER HEISER: It's not integrated at all in the way the people do their business today. MARY BETH LOUGHTON: It's a revolutionary technology. CHRISTOPHER HEISER: Certainly, the machines are getting much, much better at a lot of things than we are. MICHAL LEV-RAM: It's going to be much more integrated into our lives. CHRISTOPHER HEISER: I think we're just at the beginning of trying to understand how it fits in. [MUSIC PLAYING] 

Insights: The Promise AI Holds for Business

Uncategorized

The wonders and dangers of Artificial Intelligence.

The Promise AI Holds for Business

Fortune's Brainstorm Films explores the wonders and dangers of Artificial Intelligence. (October 2018)

now May I being incorporated into all sorts of things the Alexa Echo or Google Home or Serie? That's very garden variety. This is really first inning way. Always try to not just technology for technology's sake, but actually solving a consumer problem or main point. We feel like we need to have the right data and the right talent actually really take advantage of a I. So we're kind of focused on that, first, experimenting in a couple of different ways of getting some aye aye out there. Most CEOs were asking themselves, What can I do with a I? How can I incorporate this? This is clearly the future, but they don't really know. And that's part of the problem. It's very abstract. Most of the applications you see today involved machine learning, Google Image search, right, That's machine learning. Comparing a certain kind of add on Facebook that machine learning trying to tailor those adds to the preference is that you have expressed on Facebook the more advanced stuff that's not really in the wild yet, but for the largest tech, Artificial intelligence actually has its roots back in computing. Alan Turing, even he was really trying to recapitulate pieces of intelligence, what made a machine intelligence. And he recognized very early on that that lived in the information processing information to make it useful. Somehow term artificial intelligence was coined. In the sixties. We started to understand the brain was definitely electrical in the forties. We actually defined how it works electorally, how information is represented, what we call spiking that actually was concurrent with a lot of innovations on computing side because, he said, Well, Marines are information tools, computers, air information always parallel. First computer science department At First Neuroscience Department started a year apart. We figured out that neurons represent information certain way we tried to simulate those things and synthetic systems. We thought it would be really easy because he figured out all these principles. Brain turns out a little more difficult. So the field kind of had its ups and downs until computing got to a point where it was dense enough and the data set size has got to the point where they're big enough. 10 4011 timeframe. When neural networks met, Moore's law met big data. This whole field exploded. It's gonna change with time But overall, the goal of ah warning system is to find useful structure in data prominent Ai ai. Researchers have said that a I will be like electricity. Electricity was a big deal. It was this new invention that people talked about. Nobody talks about electricity anymore. And that's where we are with a I X years down the road sooner rather than later, eh? I will be in used in all these things in ways that we couldn't possibly battle today. The way that you used a telephone is different from the way that I used the telephone. The way that we actually came, we have artificial intelligence that can detect your behavior. So if I left my phone there and you picked it up and try to make a payment, it would know not me. When we're unloading a truck full of strawberries, one of the drivers and open up the case and take a photo waken determine whether it was sent sustainably. How long that food is gonna last. I think a I can be the future of personalization, and I think personalization is where retail and you're going. We have what's called for virtual artists which is our augmented reality beauty experience. We're using computer vision so it can kind of map your facial features and know where to place the makeup When you look at it. Virtually, we've actually been using a I really re shape the entire into end shopping journey for the customer increasingly used a I to show up through targeted display at places where the consumers shopping. So if you let us know that you're interested in certain lifestyles and certain items, we'll make sure that we personalize the recommendations that we make for you. Our customers use a lot of self driving systems. Music for object detection way work with people that are receiving the occupants. Are they happy or they're scared of losing trust. There are already dozens of application today. There is no technological development with more potential than artificially because it is a suite of capabilities. It's gonna solve simple things in our lives and the extraordinary What is really exciting about that business community? There's money to be made. There's so much excitement around, eh? I rightfully so. But if we're not thoughtful about how we go about building these systems, the outcome could be disastrous. if there aren't enough of us engaged and I don't actually even just mean women. I mean, a diverse group of people. Then we're gonna get computers who think like some bro culture on the West Coast bias ca NBI inherent to a i into algorithms if we allow it to be come down to people with those teams aren't reflective of humanity than the algorithms aren't going to be reflective of humanity. And that has hugely negative ramifications in one of the biggest concerns is that workplaces will lose their humanity if we keep incorporating more and more A i into the office. There's a lot of algorithms that are being put into place that helped the companies in the recruitment process is who they're gonna hire, who they're going to fire on so want. But the big question is what data they're using. Are they bringing in your social media data? Your private data? Do they get access to your health file your financial data, the amount of information that we freely give or unknowingly give to corporations and governments Massive, eh? I will allow those groups to manage in a very efficient way, Whether that's something that's good for me or bad for me and think depends on who's doing it and how we have all sleep walked into this current situation. We haven't really understood what itwas we were giving away. When we're using all of these service is when we're using social media. What is it we are paying with? Get a load of these. We've got people arguing that they eyes great and other people are doing the end of humanity. The answer is always somewhere in between it It's something we have to think deeply about. Every single technological development has changed jobs, right? I mean, that's nothing. You where I think this whole transformation is new, is its magnitude. It's gonna affect a lot of us. E. I'm concerned that these high discussion about evil robots it whitewashes the real challenges. And I think this is a big one. A lot off jobs would change, meaning that we will have to learn new things. And this brings us to the whole question off re and ups. Killing this is gonna be one of the major challenges of our time, making sure that everybody can follow the developments and have an opportunity to live in the labor market. What happens is jobs shift right. We have about 18,000 new jobs just for grocery picking and delivery service. Robotic. Aye, aye. Technological advances are rapidly disrupting and transforming the nature of work and the workplace. We know what happens with industrial Revolution. There is a transition time in which the whole bunch of people get disrupted right out of their jobs. Once it's through that transition time, there are better quality jobs and more jobs for everyone. Predictions are a I will create many more jobs than it will destroy. If those jobs that will be created are bad, jobs are paid under the living wage, a precarious, then that is nothing we should celebrate. I don't know if you've seen this picture of sheep problem. You showed a computer scientist three or four years ago a picture of a flock of sheep and asked them to build an algorithm to count how many people were in the picture, they would have told you is impossible. Now that's trivial. Overnight that change when I think about how fast a I machine learning is gonna move. And that's the kind of speed we're looking at. That's exciting. Right now. The depth of expertise required is extremely high. It's not integrated at all with people today. Revolutionary technology. Certainly the machines are getting much, much better. A lot of things. Yeah, I'm going to be much more integrated into our lives. I think we're just at the beginning of trying to understand.

[MUSIC PLAYING] JEREMY KAHN: One of the more frightening use cases of AI right now is the advent of deep fakes. These are videos that have either been manipulated or created in their entirety through machine learning. They've been used in some advertising campaigns. Some charities have used them for public awareness campaigns. But unfortunately, they've also been used already in cases of revenge porn. And there is a real fear that they're going to be used for political misinformation campaigns in the future. Obviously, another big concern with AI is about job displacement and the fact that a lot of skills right now that are performed by humans can be performed by machines. I know Francesca you've been doing some work on which task can be automated, you think, and which can't. And you want to tell us a little bit about that research. FRANCESCA ROSSI: Sure. So that's the work that IBM did to the context of the partnership that we have with MIT. And that work is really interesting because it uses AI itself to analyze the impact of AI on jobs. So it's another example where AI itself can be used. So we use machine learning techniques to analyze about 170 million job postings from 2010 to 2017 to see the trends on jobs. But it's important to make the distinction between jobs and tasks within jobs. Because all jobs will be transformed and some tasks will be automated, will pass from human beings to machines. So not the complete job itself, but the specific task. And of course, some tasks are easier to be tackled by machine learning approaches. So those that require the analysis of big amounts of data, the prediction of what happens in the future, and so on. Other tasks are less easy for AI and machine learning to be taken over. But some of the findings of these interesting work is that these passing of tasks from human beings to machines is happening and we see the trend. But it's not happening as fast as we thought. So this is very reassuring because it gives time to train and retrain the workforce. And we have at IBM many different programs to do that. So that the workforce is ready for the transformation. And the people can actually be ready to work with their new job, which where they have mostly to team together with AI algorithms. JEREMY KAHN: Francesca, you mentioned biases as one of the issues facing AI at the moment. And it seems like every time there's a new development, a new model people are really excited about, it doesn't take very long for someone to come along and find out that actually the data set from which the model was trained had a lot of hidden bias. There was just an example of a natural language model that Google came out with in the last 18 months called BERT. It's showing great promise in a lot of things that Google's just change its search engine to be based in part, at least in English, on BERT. But they've found that because BERT is trained from a vast amount of human generated text that's available on the internet, it exhibits gender bias, in particular because it's been trained on all of this text which has been generated over decades when people were maybe less sensitive to some of these gender issues. How do we solve this problem? I'll go to Anton first. How do you deal with this? ANTOINE BORDES: Yeah, I mean, the first step of dealing with this is recognizing that the bias is there in the data and that all data set are biased, which is something that is top of mind for most of the developers now in AI, I hope. And actually, in a company right now something that's very important is that your data set will be biased, so you need to have tools internally to check the bias to see and to try to counterbalance it. So that's the first part. FRANCESCA ROSSI: It's very important to understand also that any one of these issues, whether its bias, and we have another open source tool kits for explainability, another one for robustness. So whatever the issue that you've focused on is not something that can be even identified and resolved by AI experts alone. So it has to be addressed in a very multidisciplinary context and multi-stakeholder context-- those that develop AI, those that deploy, those that use it, and those that are impacted by the decision of the AI system-- just to understand the right notion of fairness to be used in a certain context. So we hope that these open source tool kits are raising the awareness and advancing the understanding of what needs to be done. [MUSIC PLAYING]

Insights: Confronting Deep Fakes, Job Displacement, Bias, And Other Worries Of AI

AI can be marred by bias that affects women and minorities.

Confronting Worries Of AI

AI can be marred by bias that adversely affects women and minorities. (November 2019)

way, one of the more frightening use cases of a I right now with the advent of deep fakes. These are videos that have either been manipulated or created in their entirety through machine learning. They've been using some advertising campaigns. Charities have used them for public awareness campaigns, but unfortunately they have also been used already in cases of a revenge porn. And there is a real fear that they're going to be used for political misinformation. Campaigns in the future are another big concern with the eye is about job displacement. Um, and the fact that a lot of skills right now they're performed by humans can be performed by machines. I know friends. You've been doing some work on which task could be automated, you think in which can't. And I know you want to talk a little bit about that research so that the work that IBM did the context of the, uh, partnership that we have with M. I t. And that work is really interesting because of uses a eye itself to analyze the impactor over Italian job. So it's another example where a guy can you so we use machine learning techniques that one alive. About 170 million job posting from 12,017 toe. See the trend own jobs. But it's important that you make the decision between jobs and passed within jobs because all jobs will be transformed. But on some tax will be automated will park from human beings to machines. So not the complete job itself, but the specific talk and, of course, some tax. He's here, Toby, tackled by a machine learning approaches. So those that require their knowledge is a big amounts of data, the prediction of what happens in the future. It's on other, less easy for a guy machine learning to be taken over. But some of the findings of the interesting work is that these passing off track from human beings to machines is happening. Oh, and we see the trend, but it's not happening positively fought. So these very we're shooting because it gives time, toe train and retrain the workforce. And we haven't idea many different programs to do that, so that the work even ready for the transformation and the people can actually be ready to work in with the new job, which is where they have mostly team together with a eye on Good Francesca. You mentioned bias, a CZ, one of the issues facing a eye at the moment. And it seems like every time there's a new development a new model people are really excited about. It doesn't take very long for someone to come along and find out that actually, the data set from which the model was trained had a lot of hidden bias. There was just an example of AH natural language model that Google came out with in the last 18 months, called Bert. It's showing great promise and a lot of things that Google's just changed its search engine, you know, to be based in part at least in English, on on Burt. But they found that because Bird is trained from, ah, a vast amount of human generated text that's available on the Internet, it exhibits gender bias in particular because it's been trained on all this text that has been generated over decades when people were maybe less sensitive to some of these gender issues. What? How do we solve this problem? I'll go Antoine first. How do you deal with this? Yeah, I mean, I felt the first step of dealing with it, being recognizing that the bio data and that's all that it or buy it, which is something that that stop of mind for for most of the developer 90 I hope and actually no company right night something that's very important that you know that the city via So you need to have tools internally to check the by acid to see and to try to counter balances. That's the spot, very important toe. Understand also that any one of these issues, whether it's by and we have another open source, took it for explain ability, another one for a box. So whatever the the issue that you focus on is not something that can be even identified and resolved by AI ai people, I expect alone. So it will be address in a very multi disciplinary context, a multi stakeholder context, those that developer guy does the deployed on the use it and those that are impacted by the decision of the system. Just just to understand the right notion of fairness, Toby used in a certain context. So we hope that these open source to kids are, you know, raising your Web nets and making you know, advancing the understanding of what needs to be done

[MUSIC PLAYING] You see about 70% of companies are getting into the basics of AI on some level, and maybe 15% have some of this scale. How do you make that dividing line? And what does it mean that they're getting into the basics of the technology? Well, I'd say there's an individual who's in charge of data analytics, artificial intelligence. There is a lot of pilots typically in experiments that aren't moving the needle, but they're really trying to do. So you tend to still be mostly IT-led in terms of most companies, so not really yet really led by the business side and a lot of awareness. That's where many, many companies are pivoting to now, OK, we get it. So that's the proof phase. And then you've got another 15% of companies who are in what we would call sort of pockets of greatness, right? So the comms industry, particularly in the US, for example, has now really embraced artificial intelligence in the call center area. That's the second biggest cost center after the network. And we've been working with companies where you now have 70% of all calls are digital, and what used to be 20 minutes with a human is now three to four minutes. So reducing cost and significantly raising customer engagement, and also freeing up the people in the call center to do a lot more interesting work. I was in a lab, one of our labs in Boston, and you walk into one of our traditional labs. And of course, as you'd expect, there was a lot of people in white coats with safety goggles on. Right. It's good practice. You turn a corner, and there's a robot, and the robot is processing over 200 different proteins against tissue types 24 hours a day to try and work out what could happen. And taking out, not replacing people, don't get me wrong, but replacing the capacity of the people that we had. Meaning, we can do 20x, 50x what we used to to find new medicines. [CHIMES]

Insights: How Humans And Machines Can Work Together To Maximize Productivity

A.I., automation, and analytics are transforming business. 

Humans And Machines Can Work Together

A.I., automation, and analytics are transforming the business landscape. (November 2019)

do you see? About 70% of companies are getting into the basics of a on some level on maybe 15% have some of the scale. What? How do you make that dividing on? What does it mean that they're getting into the basics of the technology? Well, I'd say, you know, there's an individual who's in charge of Data Analytics, artificial intelligence. There's a lot of pilots typically, and experiments that haven't aren't moving the needle, but But there's actually they're really trying to be, so you tend to still be mostly I t lead in terms of it. Most companies not really yet really led by the business side on a lot of awareness. That's where many, many companies are pivoting to will now. Okay, we get it, we call it That's the proof phase on. Then you've got another 15% of companies who are what we would call sort of pockets of greatness, right? So the coms industry, particularly in the US, for example, has now really embraced artificial intelligence in the call center area. That's the second biggest cost center after the network, and we've been working with companies where you now have 70% of all calls or digital. What used to be 20 minutes with a human is now 3 to 4 minutes, reducing costs and significantly raising customer engagement, and also freeing up the people in the call center to do a lot more interesting work. I was in a lab, one of our labs in Boston, and you walk into one about traditional labs on. Of course, as you'd expect, there is a lot of people in white coats with safety goggles on, right. It's good practice. Um, you turn a corner on dhe. There's a robot on. The robot is processing over 200 different proteins against tissue types 24 hours a day to try and work out what could happen and taking out. Not not replacing people don't get me wrong, but replacing the capacity of the people that we had meeting weaken do 20 x 50 x What we used to find you

[MUSIC PLAYING] RICHARD SOCHER: And AI is not dual-use but omni-use technology. In the end, we don't say, oh, what about computers? They can be used for good and bad. Yeah, computers can be used for anything. And it gets even simpler, like, a hammer. You can use a hammer to build a house. You can get a hammer to hurt people. You can use the internet to do great things. You can use cars to do-- AI will be everywhere. Every single industry will be impacted by AI. And in the end, it's up to us to define what tasks we want it to do. And then when we define those tasks, we also need to think about the data sets. AI in the end is only as good as the training data that we give it. And humans have to think carefully about ethics. And I think the field of AI has now woken up to that. DARIO GIL: So I think one of the ways that we've got to address this is to put the responsibility back to who is creating this software and what is their purpose, and what is their intent. And the accountability needs to rest with the people and the institutions that are creating this software. And we should really avoid the tendency to say as if AI as an external agency was doing this on its own. Fairness, as an example, like if we look in the context say of banking, and you're trying to build a predictive model to be able to assign credit as an example, it's based on learning from past examples. So if you just naively went and you say, well, what credit approvals have we done over the last 20, 30 years? And let me use the approvals that were generated as a means to train this model. Well, what would happen most likely is that you'll end up with a model that is likely to give more credit to men then to women. And that's not necessarily that the neural network is biased. It's the data sets that we're done were biased. So a way to compensate for that issue is you got to have robust AI engineering methodologies to be able to test for unwanted bias. And particularly, for example, if you have protected variables that are protected by law that you can not use as a means to do discrimination from the perspective of assigning classes. You just have to be able to detect that. So we believe explainability, fairness, robustness, some of these AI models are subject to adversarial attacks. And you got to have software to be able to test it. Five years ago or so on, it was just mostly the realm of a relatively few large players and of course, specialized startups that were in there. Now you go and back to the AI like software, you can go to any financial institution, insurance companies, health companies, automotive, and they want to be co-participants and co-creators of AI. So a trend that has happened has been to democratize and give tools to people, so that they can create their AI efficiently, and how you debug it, and how do you create AI engineering as a discipline just like we have done in the past with software engineering. And we will create our program, and debug it, and so on. We need all of that tooling to allow everybody to also co-create and integrate with each other. 

Insights: Who's Controlling AI?

IBM's Dario Gil on the many concerns people have with AI. 

Who's Controlling AI?

Dario Gil of IBM Research, and Richard Socher of Salesforce, discuss the many concerns people have with AI. (July 2019)

I is not really but on the use technology. In the end, we don't say, Oh, what about computers? They can be used for good about your computer's confused for anything, and it gets even simp like a hammer. You can use a hammer to build a house with a hammer to hurt people. You can use the Internet to do great things. You can use cars e. I will be everywhere. Every single industry will be impacted by eye, and in the end, it's up to us to define what tax we wanted to do. And then when we define those tasks, we also need to think about the data set. A. I in the end is only as good as the training data that we give it. And humans have to think carefully about ethics. And I think the feel if the eye has now woken up to that. So I think one of the ways that we gotta address this, I put the responsibility back to who is creating the software on DDE. What is the purpose and what is their intent on day accountability needs to rest with the people on the institutions that are creating the software, and we should really avoid the tendency to say Aye, aye, as an external agency was doing this on its own fairness as an example, like we look in the context of banking on you're trying to build a predictive model to be ableto find credit as an example you know, is based on learning from past examples. You just naively went and you say, Well, what credit approvals have we done over the last 2030 years? And let me use the approvals that were generated as a means to train the model. What would happen most likely is that you'll end up with a model that is likely to give more credit to men and a woman. And that's not necessarily that the neural network in bias is the data science that were done. We're biased. So I'm way to compensate for that issue is you gotta have robust engineering methodologies to be able to test for unwanted Maya and particularly, for example, if you have protected variables that are protected by law that you can not used as a means to do discrimination. From the perspective of a signing classes, you just have to be able to detect that. So we believe explained ability, fairness, robustness These some of these models are subject with personal attacks and gonna have software to be able to test it. Five years ago or so on, it was mostly the ram off relatively few large, large players and, of course, specialized startup that were in there. Now you go on back to the eyes like software. You can go toe any financial institution insurance company's health care companies out of motive, and they want to be poor participants and co creators. Off I sew a trend that have happened has been to democratize and give tools to people so that they can create their efficiently on how you debug it. And how do you create a engineering, a discipline, just like we have done in the past with software engineering, right, and we'll create a program and debug it and so on. We need all of that tooling toe, allow everybody to also co create and integrate with each other

[MUSIC PLAYING] When we surveyed Fortune 500 CEOs, one of the questions we ask them, was to the extent you're using AI, are you using it-- are you using business process automation? Yeah, I'm sure they all said yes. Or are you changing the value for your customers? 69% said business process automation. That is 100% replacing people. Yeah. So I said the word augment. Right? So it doesn't mean that-- and I'm not trying to be word smithing. For as long as all of us would know, technology has replaced some things. That's going to happen. But augment, if the purpose of these technologies is to let you do what you do best. What man does best, and help him in doing his job, and I have come to see that now when I've watched this stuff fully implemented at clients. Look, I think the biggest issue with AI right now, I call it that most people have random acts of digital everywhere. And they've put this sprinkle dust of AI, and they're getting no value from it. Because it's just little bits of things. They didn't really change how any work got done. This is kind of chapter two. There's some pivots coming here, where the movie's going to go. But that idea that you're augmenting, that's why-- if you ask me, like some of my own lessons learned in implementing in IBM. Probably the biggest set of skills I had to build were all human centered skills. They weren't teaching people like the technologies of AI. It was having to put in 20,000 design engineers, 50,000 studios, 200,000 agile. Or I couldn't get the value out of putting the technology in the middle of a new workflow. But the business process automation is the easy stuff. That's the low hanging fruit. If you want to just do what you used to do automated. Most people don't want to-- [INTERPOSING VOICES] That doesn't augment humans. That just replaces them. Yes. So how do you get people to focus on where the value is? Not the easy low hanging fruit. Or not just the easy, low hanging fruit. Well, look, I think we're entering kind-- I view a chapter two here now. What people like, to date, almost all digital transformation-- everybody would say digital is the reason they're transforming. And I think, to date, everyone would say the bar has been set by born digital, born on the cloud, set the expectations of where experiences. I think you're kind of entering a second chapter. In this first chapter, most people I know who did anything with AI and automation, it was very much outside in. It was let's put something new out for customers on the front end. Something new. An outside in application. Typically customer facing. New and additional. And I think, when you say, how do I get this to go for the good? You were additive in what you were doing. You didn't fundamentally change how you did claims processing. You didn't fundamentally change these things yet. And so, I think we're about to start that side now. So I don't think people have seen that. We're in the early inning. I think we're in the very early innings of this. And the reason why, I always say, one of my biggest lessons learned on deploying AI, the hardest thing is changing the workflow and the change management. Because you've got to change how people do work. That is really the hard part. And we're just now coming into it. [MUSIC PLAYING] 

Insights: Ginni Rometty Predicts Rise of Cognitive AI 

The IBM head spoke at CEO Initiative 2019.

Ginni Rometty Predicts Rise of Cognitive AI 

when we surveyed Fortune 500 CEOs. One of the questions we ask him, What's to the extension using a r A. I Are you using it? Are using business process automation? Yeah, I'm sure they said yes. Or are you changing the value for your customers? 69% said business process automation that is 100% replacing people. Yeah, so I said, the word augment, right? So it doesn't mean that I'm not trying to be worth missing for as long as all of us would know. Right? Technology has replaced some things. You know that that's going to happen. But augment. If the purpose of these technologies would let you do what you do best, what man does best and help him in doing his job, and I have come to see that now, when I watched this stuff fully implemented it clients. Look, I think the biggest issue with a I right now, I call it that most people have random acts of digital everywhere, and they put the Sprinkle dust of a I, and they're getting no value from it because it's just little bits of things they didn't really change. How any work got done. This kind of chapter two. There's some pivots coming here where the movie is going to go. But that idea that your augmenting death white if you asked me like some of my own lessons learned in implementing an IBM the biggest set of skills I had to build. We're all human centered skills. They weren't teaching people like the technologies of a I. It was having to put in 20,000 design engineers 50,000 studios, 200,000 agile, or I couldn't get the value out of putting the technology in the middle of a new workflow. So but the business process automation is the easy stuff. That's the low hanging fruit even wanted to do it used to do automated. Most humans just replaced. Yes. So how do you How do we How do you get people to focus on where the value is? Not the easy, low hanging fruit? Yeah, or not just the easy lowing Well, look, I think we're entering kind of eye view a chapter to hear now what people like to date almost all digital transformation. Everybody say digital's reason. They're transforming. And I think to date everyone would say uh, the bar has been set by born digital. Born on the cloud set the expectations of where experiences I think you're kind of entering a second chapter in this first chapter. Most people I know who did anything with a I and automation. It was very much outside in it was, let's put something new, uh, for customers on the front end, something new on outside in application, typically customer facing new and additional. And I think when you say, How do I get this to go for the good? You were additive in what you were doing. You didn't fundamentally change how you did claims processing. You didn't fundamentally change these things yet. And so I think we're about to start that side. Alice, I don't think people have seen that. So we are in the early and I think we're in the very early innings of this, you know, and I and I in the reason why always say one of my biggest lessons learned deploying a I. The hardest thing is changing the work full in the change management. You got change of people to work. So this is that is really the hard part, and we're just now coming into it.

[MUSIC PLAYING] DARIO GIL: I think what's happened right now is that AI has really begun to come into our lives. And we've begun to see devices like a smart speaker where now we have a speech interface. And underneath there will be AI models that are implementing the speech recognition and then we can do simple functions. Or we may have in our cars some form of intelligent assist. To give one last example, better weather model prediction. Is it going to rain? Or will they have also applications in yield prediction in agriculture? We've seen the increased use of machine learning models behind the scenes. So in many different aspects of our lives, a narrow form of AI has begun to work and be present. One of the most exciting ones in the field of AI is something called Project Debater. So we have been working for the last six years within IBM Research on creating an AI system that can engage in a full live debate with a human being from one sentence-- not one sentence, one comment, one sentence, 10 minutes of content that also involves some amount of the listening, comprehension, and the rebuttal. That to me is an amazing achievement that is indicative of what is the next phase for AI, going from a narrow AI to a broader form of AI that is more capable and more sophisticated in the use and understanding of language. If we paint the broadest possible picture, we can envision a future in which we humans, we're going to have a complementary expertise, as we go on tackling making decisions. Today, we form institutions, companies of different forms, where through the expertise of the people who work there, we work with each other collaboratively to solve problems. Why couldn't we have a form of AI that is part of those colleagues? Definitely, we're making advances in making AI easier to use. One key area is that as we make more progress on technologies like transfer learning and bringing in reasoning along with learning, the amount of data that is required to train the system goes down. A second one is around the automation of AI. It is becoming easier and easier to have machines to automatically construct both the neural network and increasing amounts of the data science associated with training these systems as well. The computation for AI is also improving quite dramatically. So you can train more efficiently at lower cost and faster. We are going to see increasing democratization of the ability to create AI, and very importantly, the ability to deploy AI, and make sure that it is trusted, which to me is the big, big new 2019 center stage is going to be the issue of not just accurate AI, but trusted AI. As we think about AI, learning is not enough. We also need reasoning in addition to learning. Because it allows us to create also models of the world and think about possibilities we haven't yet encountered or for which data is not there. When you actually look into the details of what that would mean, it is likely to be the basis of a significant leap in reading comprehension. And if we could actually build AI systems that comprehended text, and books, and so on, that would be a significant and very important milestone for the field that would utter the advances of creating these systems that can really complement our knowledge and our expertise. Just like in the Industrial Revolution, we extended what our muscles could do through machinery. Now, we would truly have a companion to help us expand what our knowledge, and our reasoning, and our decision making can be, but it starts with unifying learning and reasoning. [MUSIC PLAYING]

Insights: IBM On How AI Is Already At Work

IBM's Dario Gil on what we'll see in 2019 and beyond.

IBM On How AI Is Already At Work

Dario Gil, IBM's VP of A.I. and quantum computing, breaks down what we saw in 2018, and what we'll see in 2019 and beyond. (December 2018)

I think what it happened right now is that a I has really begun to come into our lives on. We begun to see devices like Smart Speaker, where now we have a speech interface and underneath there will be a I models that are implement in the speech recognition and then weaken the simple functions. Or we may have in our cars some form of intelligent assists to be one last example. Better weather models prediction. It's gonna rain or it will be handled. The implication in like your prediction in every culture, we seem to increase use machine learning model behind many different aspects of our lives. A narrow form of a I has begun to work the press one of the most exciting ones in the field of a I from single project debater. So we have been working for the last six years within IBM research on creating a system that can engage in a food life debate with a human being for one cent, not 1/10 of one command, one sentence, 10 minutes of content that also involved some amount of like the listening comprehension under a bottle. That is an amazing achievement that is indicative. What is the next Phase II going from a narrow way, I tow a broader formally I that is more capable and more sophisticated use of language, way paint, their broadest possible picture. We can envision a future in which we humans we're gonna have a complimentary expertise as we go on attacking, making decisions. Today we form institutions, companies of different forms where who with expertise off the people who work there. We work with each other collaboratively to solve problems. Why couldn't we have a form of the eye that is part of those colleagues? Definitely. We're making advances in making a I easier to use. One key area is that as we make more progress on technologies like transfer learning on bringing in reasoning, along with learning the amount of data that is required to train this system goes down. The 2nd 1 is around. The automation of it is becoming easier and easier to have machines to automatically construct both the neural network and increasing amounts of the data science associated with training. These systems as well the computation for a I. So it's improving quite dramatically so you can train more efficiently at lower cost and faster. We're gonna see increasing democratization off the ability to create a I and very importantly, the ability to deploy a make sure that it is trusted, which to me is a big, big new 2019 center stage is gonna be the issue of not just accurately I trusted me. I way think about a I learning is not enough. We also need reasoning in addition to learning, because it allows us to create also models of the world and think about possibilities we haven't yet encounter or for which data is not there. When you actually look into the details of what that would mean, it is likely to be the basis of a significant lead in reading comprehension. I mean, we could actually really I system that comprehended tax on books and so on. That would be a significant, very important milestone for the field that would utter their advances. Off created systems that can really compliment or knowledge or expertise, just like in the Industrial Revolution. We extend that what are muscles could do machinery. Now we would truly have a companion to help us expand what our knowledge reasoning on our decision making candy. But it starts with define learning and reasoning

 [MUSIC PLAYING] MARISSA MAYER: What impact do you think AI is going to have on the workforce both in the short term and the long term? MIKE ABLESON: So it's a topic that obviously, we could spend a whole conference on. As I said, in the short term, I expect that it's going to give us a whole lot more tools to do our jobs better. And it'll be an extremely, competitive space on how companies adapt and then deploy AI. Longer term. It is going to change the nature of jobs. And you know, we should be justifiably concerned about that and thoughtful about it, about what we're going to do about it. I think we have time to be thoughtful over the next few years on what we're going to do. Those people that are doing jobs that aren't going to be the same jobs going into the future? We need to think about what do we do to help them move to something else where they can have a new job. MARISSA MAYER: (INHALES DEEPLY) MIKE CAPPS: Yeah. MARISSA MAYER: Any thoughts on the--? MIKE CAPPS: I'm wildly uneducated in workforce retraining. I just know that the number one job in North Carolina, where I'm from, is truck driving, and also in California, and also in Texas. So that's a big problem, and I don't think it's 30 years away, right? It's a lot sooner than that. What I'm more excited about is workplace support, finding ways to remove the cognitive bias from boring tasks, from repetitive tasks, from really important tasks that you probably shouldn't be making the decision about what to say on this panel, when drunk the night before preparing for it. Right? These are things you shouldn't do. And if you've got an AI to help you with that, maybe, it'll save you from making some mistakes when you have your best you available all the time. [MUSIC PLAYING] 

Insights: Artificial Intelligence Will ‘Change the Nature of Jobs’

Artificial Intelligence Will ‘Change the Nature of Jobs’

We should be concerned and prepared. (July 2018)

What impact do you think I'm going to have on the workforce? Was in the short term and long term. So it's a topic that obviously we spent a whole conference on, and they said in the short term, I expect that it's gonna give us a whole lot more tools to do our jobs better. And it'll be a extremely competitive space on how companies adapt and then deploy II. Longer term, it is going to change in a nature of jobs, and, you know, we should be justifiably concerned about that and thoughtful about it about what we're going to do about it. Um, I think we have time to be thoughtful over the next few years on what we're going to do, and, um and we need to, you know, those people that are doing jobs that aren't going to be the same job's going to the future. We need to think about what do we do to help them move to something else where they can I have a new job? I'm wildly uneducated and worked for three training. I just know that there's the number one job in North Carolina where I'm from his truck driving and also in California and also in Texas. That's a big problem, and I don't think it's 30 years away, right? It's what I'm more excited about is workplace support finding ways to remove the cognitive bias from boring tasks from repetitive tasks from really important task that you probably shouldn't be making the decision about what to say on this panel? Win drunk the night before preparing for it right? These things you shouldn't do. And if you've got an A I to help you with that, maybe it'll save you from making mistakes. When you have your best, you available all the time.

Insights: Artificial Intelligence

Insights: The Promise AI Holds for Business

Latest video

Fortune's Brainstorm Films explores the wonders and dangers of Artificial Intelligence. (October 2018)

Videos (7 videos)

Insights: The Promise AI Holds for Business

Insights: Confronting Deep Fakes, Job Displacement, Bias, And Other Worries Of AI

Insights: How Humans And Machines Can Work Together To Maximize Productivity

Insights: Who's Controlling AI?

Insights: Ginni Rometty Predicts Rise of Cognitive AI

Insights: IBM On How AI Is Already At Work

Insights: Artificial Intelligence Will ‘Change the Nature of Jobs’