Artificial intelligence and algorithms: pros and cons | DW Documentary (AI documentary)

Artificial

intelligence

is advancing rapidly, there is talk of a new evolution that could fundamentally change life on our planet. Artificial

intelligence

has the potential to revolutionize all aspects of daily life, labor mobility, medicine, the economy and communication. but will it really make medicine better and doctors superfluous when Will self-driving cars hit our roads? Will intelligent robots usurp our jobs? Are we heading towards a dystopia without privacy and total surveillance? What exactly is

artificial

intelligence and how much can it really do? What will change and what will remain as pure fantasy to answer. these questions we embark on an exciting journey to meet the scientists working on our future in the USA, Great Britain, Germany and China, our first stop in Silicon Valley in California, Apple, Google and Facebook are headquartered here.

It is the epicenter of the digital revolution that the technology industry has had. changed the face of the san francisco bay area new businesses launching every day rents have skyrocketed and

artificial

intelligence is the buzzword a new type of supermarket recently opened its doors here amazon go all you need here is an app Bring your mobile phone to the scanner and you're in while Leonardo shows me the new Amazon menus and explains that the language assistant Alexa can help with preparation at home I'm under

cons

tant surveillance which shelf I stop at which products I'm interested in on the ceiling sensors and cameras Intelligent image recognition captures my every move.

What do I take off the shelf? What do I leave behind and what do I take with me? This branch is still in its testing phase, but Amazon plans to open 50 such grocery stores this year alone, by the end of the year. sales assistant just walk out I don't have to wait in line anymore there are no cashiers I feel a bit like a thief as I leave comfort at the cost of privacy my receipt a block away a robot cafe another testing lab for the future order by app and touch screen increasingly ubiquitous commerce tools my first cup of coffee served by a robot so this is the taste of the future AI will change our shopping experience, but what will happen to employees?

Stanford University is at the forefront of global AI research with an annual budget of $6.5 billion I want to know how artificial intelligence will change in medicine researchers here have developed an artificial intelligence algorithm that can detect lightning X to detect certain diseases, computer scientist Pranav Rajpurkar shows me how easy it is to use taking a photo of an possibly cancerous and I can see it right here, okay, so I'll give it a look if I can take a look. Now there's a chance of pneumonia, inaudible edema, stroke and that goes to the bank, yeah, now how does this work?

I mean, how did you get to the point where we started with a large data set of chest x-rays that were published by the NIH and that contained X-rays and then also labels of different pathologies and whether they existed on those x-rays, so you could say well, here is an image and in this image I have pathology one, two and three and we had hundreds of thousands of these images, so we trained a model that can take it as input and x-rays and then generate the probability of various pathologies different in these x-rays. Artificial intelligence is modeled on the human brain.

A gigantic network of almost 100 billion interconnected neurons, put in very simple terms, this is how a brain cell works the incoming impulses are passed in a domino effect from one neuron to the next the resulting circuit connects the neurons and is this circuit which artificial intelligence tries to simulate as a digital neural network like our brain, the network can learn to identify tuberculosis, for example, first the network needs to be trained or taught, X-rays of tuberculosis patients are sent to the system initially, it It is difficult to correctly identify the condition, but every time an x-ray is fed, the network structure adapts and its diagnostic ability improves Thousands and thousands of clinical data sets are needed to train the machine only after the network is optimized So, can you correctly identify an unknown x-ray, but how accurate is artificial intelligence compared to a doctor's experience?

In fact, we have performed this test twice. At this point, once with a set of studies from the NIH data set we had a group of radiologists labeled and then compared the accuracy of the model to that of the radiologist and found that they were very similar in terms of accuracy in most of the the pathologies. in one of them the model outperformed the radiologists and in three of them the radiologists outperformed the model and then we repeated the experiment this time using a Stanford data set that we recently published, which is 200 thousand chest x-rays and then we had a similar setup where we had three subspecialty radiologists, these are very rare and highly trained radiologists, to decide what the ground truth was for a particular set of images and then we compared the general radiologists to the task algorithm and found out that had similar levels. of performance, so these are all radiologists from Stanford, so they are trained, they should be good, yes, reading x-rays accurately is a complicated process, but artificial intelligence is progressing rapidly when it comes to identifying or recognizing simple images that computers have surpassed. human precision now, if I look at your photo, it's always probability, so there are cases where the machine is not really sure what would be a clear decision to say "okay", this is "I don't know, pneumonia or something else ", Yes I think so.

I mean, I think it's good to talk in terms of probabilities because probabilities also give insight into the algorithm, the uncertainty of the model in that particular problem, I think one difficulty with probabilities is that it makes it difficult for humans to interpret which is the probability of 88 versus 92 percent in terms of the decision that I have to make in the clinic and therefore I think in that sense one of the things that we could experiment in the future is, instead of showing probabilities that are so detailed, that perhaps we can. show things as improbable or this pathology is probable or this pathology is possible in healthcare artificial intelligence is driving a revolution, scientists are using artificial intelligence

algorithms

to examine seemingly banal data, such as movement up and down of the steps we take every day.

We are looking for striking patterns that can serve as early warning signs of the disease. Scientists in the English city of Birmingham are working on a revolutionary diagnostic method. To date, there are no specific tests to detect Parkinson's disease, making diagnosis difficult. This could change that. mathematician from Aston University very good voice changes can be an early indicator of Parkinson's peak and his team collected thousands of vocal recordings and fed them into an algorithm they developed that learned to detect differences in voice patterns between people with and without the disease in a laboratory. Based on a study of the recordings, the algorithm was able to correctly identify the Parkinson's diagnosis almost 99 of the time.

Max Little's work is an example of the far-reaching changes that AI is bringing to the field of medicine. It is no longer just doctors who use artificial devices. intelligence to develop new diagnostic methods, but data scientists, programmers and mathematicians like at most one example when a person walks, the sensors in his smartphone record the movement up and down of his door, but what Information can be obtained from that data if we measure someone's walking pattern? behavior then someone who is healthy could have it, we could measure the accelerometer so it looks like this, okay, so that's the type of movement you would have here, yes, the hips go up and down regularly, that kind of thing along with your rhythm, but if you look at someone with parkinson's disease they can have these little steps like this and they can be irregular or they can have patterns like that or they can even freeze and stop like this so you can see there's a difference there's a difference so you can Now We also train an algorithm, for example, to select features like what is the distance between these peaks and you could also do the same thing with this and you would be able to do it very accurately and by doing that, you might be able to measure, for example, that there are a great variability between these.

The advantage of the algorithm really comes when, for example, you can have someone say who measures a pattern that looks like this and there might be only one. small change maybe that happens very, maybe not like that, but in a way you know some more variation that is in the sequence of these um at the time of these events um even to a professional eye because they don't have the level of precision that is they may not be able to detect that this is outside the normal range of um a variation but of course an algorithm connected to a high precision sensor um you know we'll be able to determine that. difference and in this case this person here may in fact have a precursor symptom of the disease, so this would mean that this person with the help of an algorithm could be diagnosed with Parkinson's, while the doctor himself would ignore it, which It might, for the moment, make it possible to detect the precursor symptoms of Parkinson's and enable early intervention, but what else does the data from our smartphones reveal at the moment?

You already have apps that track your supposed activity. Yes, in fact, the data may already be there. It could potentially be there, that's right, but there's an ethics to whether we collect that kind of data and use it for these kinds of purposes. You clearly know that we can't just collect this data and start diagnosing people, which we certainly shouldn't do, at all. We could, but we wouldn't really want to, there are very good reasons not to do it and there may be good reasons to do it too, but those are the kinds of things that need to be resolved in a uh in a Well, you know, in a regulated environment afterwards From our interview, Max Little tells me that he has received several lucrative offers to join tech giants that smell new business opportunities.

He rejected them. Artificial intelligence will undoubtedly improve doctors' ability to detect and diagnose diseases, but amidst all the opportunities that AI offers. an urgent need for regulation we are on our way to china a country that has undergone impressive changes in recent years its capital beijing is bursting at the seams the entire country is hungry for progress and is on the fast track to the future time seems to move faster here By 2030, China aims to be the world leader in the field of artificial intelligence and there are many indications that it will reach that goal because the government has funded subsidy programs worth billions of euros.

These robots don't assemble cars, they are the big attraction at Beijing's latest smart restaurant ai in the kitchen and automated waiters I have a meeting here with design researcher Geisha Yost former Internet ambassador for the German government currently spending a semester of research At Tongji University in Shanghai I asked him about his impressions. In China there is real hunger in the city and it is very fun to talk to young people because they want to be the drivers of change. They work day and night. They have a new work-life balance model called 996. I was like, what are you doing? means 996. and they said we work from 9 am to 9 pm six days a week, that is the best model now because they used to work non-stop but no one stops, no one slows down, they work like crazy because they want to make a change.

Since this restaurant costs 20 million dollars in one restaurant alone, they have invested this huge sum to digitize the entire operation. They are not just robots that serve food. The entire kitchen is digitalized. Refrigeration is monitored. Supply chains are monitored. There are control panels for everything. Everything is connected here where they are testing what works and what aspects can be implemented in other restaurants of this chain. That's the idea here, just try things and think big, but what about privacy? There seems to be a trade-off between security. and privacy, you often hear how AI has increased public safety, for example, that ubiquitous surveillance cameras have dramatically increased the crime solving rate, making it difficult for us to relate because privacy andPersonal rights are very important to Germans, but there is a different tradition and take here.

On the subject, I am fascinated by China, but I am also baffled how the high civilization of ancient China and the modern industrial state can be reconciled with surveillance cameras everywhere in the Long Gan district of Shenzhen, in the heart of the burgeoning region. China's economy north of Hong Kong we visit the smart city control center a giant monitor shows data from the entire district in real time numbers of new residents by neighborhood to plan schools water supply levels power outages all this Information is collected, compiled and evaluated using artificial intelligence. The exhibition project was developed with Chinese technology giant Huawei's chief engineer, Chen Bangtai.

He tells me that the city now operates more efficiently, so what are you doing here? It is urban planning. Yes, the systems are very helpful. These are hospital beds. There are currently 15,000 doctors and nurses. 7,600 beds, so Shenzhen is currently healthy or sick. An intelligent surveillance system scans the entire city. Illegal structures like this one on a rooftop are quickly identified and demolished. Some of this seems like the backdrop for a science fiction movie to me. Employees with live streaming. Cameras survey side streets. This is total surveillance. Chan shows me how cameras are installed in restaurant kitchens.

He even controls the cleanliness, but the chef doesn't mind being monitored all the time. Chen says residents of the Long Gang district approve that jaywalking is not allowed and that violators are identified immediately. Look here, you jaywalk once and immediately your social credit score drops to this degree Surveillance is unthinkable in the West, but here in China they have a different view and say it has caused a drop in crime. What does it say here? Young men without glasses. Youths. Yes, suddenly you are a young man. A transparent society for the sake of efficiency. This seems useful, but do we really want to measure control and analyze everything just because it is technically possible?

Won't it inevitably lead us down the path of data dictatorship? Maybe trust is better than intelligent control. Silicon Valley, synonymous with innovation and unlimited freedom. The oldest. Players in the AI ​​field are based here, but their headquarters are hidden behind discreet low-rise buildings. Facebook we use their services, we trust them with our data, but the company is impermeable to the public. A selfie at the front door is almost tolerated in the house next door. At Apple, the visitor center's 3D campus model is the closest non-employees can get to the new building. What happens inside is all confidential.

We want to visit Google here in California and requested an interview weeks before our arrival, but all we got. they are delaying tactics like these visitors google leaves us out apart from a small shop this is the only prominent visitor accessible to the public these android lawn statues are even a designated location on google maps welcome to google google the european union gives google a $2.7 billion antitrust fine These companies have increasing power over our daily lives and increasing political influence. Google spends more than 6 million euros a year on lobbying in Brussels alone. The EU Transparency Register lists more than 200 meetings with Google representatives since 2014.

Google is the busiest lobbyist in Brussels. We finally got our interview not in California but in Munich Germany with one of the longest-serving employees Jens. redma how important artificial intelligence is to google ai is so important to us that two years ago we renamed our entire research division to google ai AI drives a significant part of our product development. AI, above all, drives an important part of our efforts to improve the quality of our products. We use automatic translation through the use of automatic

algorithms

. We have seen faster progress in the last two years than in the rest.

Without a doubt, the entire society of the previous decade will be propelled forward thanks to the implementation of these services and the use of AI in the coming years. The key is that it is done responsibly under the principles of transparency. We need to explain how things work. Why are they needed? people's data goes how they can control it how they can delete it if they want to delete it or forward it the user should be in control, but what about technologies like Google Home, the smart microphone that sits in people's living rooms? Google Home is not eavesdropping, there is a small chip in the device that listens for the so-called hot word, it is waiting for the command ok google or hey google and only then the microphone is turned on to send a voice command or search request to The Internet, the Google server that then presents the results, as a science journalist, I am naturally curious about the future.

There is this patent application from September 2016. Google's application offers a detailed description of what can be inferred from household noises. How long do we brush our teeth? If we argue or if a housemate. It's sick, it's much more about capturing atmosphere and habits than words and it's a Google patent application that anyone can search. I don't know anything about this particular patent application. We have a whole series of patent applications every year, most of them are imaginary fictitious ones. services that, as in many other companies, never translate into real services services Google's lobbying activities in the EU are at least indisputably real how much does Google intervene in the EU I think the most important question is the one between the lines, that is, to what extent does a The company deals with product development and we have established our own rules in accordance with principles that guide our own actions.

Research and product development that also guides our business decisions. On its own turf in the United States, Google faces growing political pressure in Washington. We meet Barry Lynn. head of the think tank open market institute warns of the dangers posed by technological giants influence that we need to know in our society that the people who bring information to the public sphere who speak to the press who speak to the representative our representatives uh in congress who They represent themselves that they speak on their own behalf and not on behalf of another person that they are not puppets that they are not puppets and the fact is that today our society this is true here in Washington it is true in Europe our society is full of puppets with puppets representing the interests of google, facebook and amazon, given the monopoly power of big tech, the calls for regulation are getting louder in washington when you have a monopoly, whether on retail, whether whether it's about searches, then it means that the public doesn't really have the ability to understand how that information is used, how that power is used, monopoly per se, unless it's closely regulated by the public, is a danger, Google It wants to take over the world, it wants to direct our thoughts between people. and person our communications between person and person our dealings and business between person and corporation they want to direct everything they can they want to know what is happening in our thermostats and our homes they want to know what we are watching on television they are at a level of arrogance that not even not even the Stalinists could have imagined uh uh boosting google facebook amazon the influence of the tech giants will continue to grow what can be done to control their monopolies one thing is clear artificial intelligence is

cons

olidating its control over There is an urgent need to rethink antitrust policy.

Mobility is another area where AI is driving the advancement of innovation. In the near future, driverless cars could be put on city streets, but how realistic is this vision that we have come to Boston with? to the prestigious Massachusetts Institute of Technology Certus Caraman is a leading expert in the field of autonomous vehicles he and his team are working on prototypes of autonomous vehicles I think we have made a couple of points with computers and machines, one is all this mapping and location, all that technology works super well, computers can know where they are in centimeters, maybe sometimes with millimeter precision, much more than what is required to drive, computers can't look around and understand where things are. others, but that is not what is required.

To drive, what it really takes is understanding what's going to happen next in the next three seconds, five seconds, the next minutes, maybe even sometimes the next hour, and that's the key piece that's missing and I think that The problem is that right now it is very difficult for you. Describe to me how you understand whether or not a person is going to use the sidewalk or use the crosswalk and cross the street. Sometimes you look at a person's face and that facial impression gives them away and sometimes you will slow down. no, they may be looking in the same direction, they may be standing in the same place, it's just a small impression of the face, maybe just the way they are and, unfortunately, that kind of intuition, instinct, etc., is very difficult for us to program on computers.

It works in a simple lab environment, but in real-life environments the algorithms are still totally out of reach, which doesn't bother advertisers. Our test drives were nothing more than a series of failures, one inexplicable emergency stop and another on the second attempt. The sensors were activated. this vehicle was overloaded by a car parked on the sidewalk and here this smart car missed a car that swerved into our lane and which did not work. Talking to MIT engineers, it becomes clear that to build an autonomous car developers must comply a massive scale of technical requirements what I think about fully autonomous cars is that I think I would be very surprised if this happened in less than 10 years.

I would also be very surprised. I am a great believer. I would be very surprised if it happened. It's not going to happen in 20 or 30 years, I think it will happen at some point, but I do think that people really underestimate the kind of technology that needs to be built to make your car completely autonomous in all conditions, in all circumstances, in any circumstance, that is the very difficult part. Ultra fun driving is not as trivial a process as you might think and that is because you constantly have to observe what is happening around you, cyclists and pedestrians, sometimes you have to doubt whether this guy wants to cross the street or not, it is difficult imagine all that.

When calculated automatically, an autonomous vehicle would also have to be able to deal with all of this. Here we have a truck doing a three-point turn. You might have to back off now if he doesn't make it, do you want to cross? whether or not it is on the road some people don't even expect the fully autonomous car is a distant dream, but driver assistance systems are already making our roads safer an accident filmed from a car equipped with emergency braking assist the sequence of events can be evaluated in slow motion the red vehicle in front misses the approaching traffic jam, no brake lights appear, but the distance sensor of this car registers the interruptions of the traffic jam and prevents a new collision, But what principles should guide the decisions made by technology in an accident situation in recent years?

The mit media lab has been addressing the ethical questions raised by artificial intelligence, which moral compass future smart devices should refer to, and is one of the world's leading experts on these issues. He and his team developed a survey called the Moral Machine to explore ethics in programming autonomous vehicles as in the case of In an accident, most of the time people don't remember anything and don't have time to react. Everything happens very quickly so they are just surprised, maybe they see something in front of them and just veer off in some random direction or maybe they just get scared. and hit the brakes, so you can't expect a human being to do the right thing in such a short time, exactly in that, on such a small time scale, unless you know they made a decision beforehand, like knowing if they drank and they drove or they knew it.

They were going to cross a red light, so you blame them, but otherwise you can't blame the human, but a machine because of the speed of electronics, because the self-driving car is evaluating the environment, you know it millions of times per second and then time. it goes much slower for the machine and is able to recalculate the situation and maybe recalculate the strategy and this is where we canmake a potentially better judgment than any random choice the human used to make in this situation, now what is better is a very interesting option and it is not obvious and let's look at a case where we have people versus people, so now we have, the vehicle has two people in it, okay, and it will swerve and hit the barrier, so people will die in the car. or the car will go straight and kill the pedestrians, the pedestrians who are crossing illegally but they are also women and these people in the car are men, so now it becomes very complicated very quickly whether to prioritize women over men or all. be the same whether you prioritize pedestrians over passengers or not, whether you should take into account the fact that people are crossing illegally in this case, as you can see, once you have multiple dimensions, it's not obvious what the right thing to do is a do a or b who should die, the old lady crossing red or the boy in the self-driving car what choice should the algorithm make ead's online survey presents respondents with several scenarios, each with its own unique dilemma.

Respondents are then asked to choose how they would want an algorithm to decide that, as a result we have 40 million decisions and they are still being counted from people around the world and it allows us to start analyzing what people agree on, but also how they differ, and our culture influences our moral judgments. We always agree on saving more lives, saving children, saving people who cross legally instead of people who don't cross legally, and the most interesting part is that you could choose a country like Germany, yes, and you could see how they compare. with the world average.

Look, okay, so the status is not really important, but what you can see in Germany is preferring it in action, yes, so if you don't have to, if you prefer to go straight, yes, which is the default, don't make a decision exactly like that. it means that Germans don't like to make a decision, yes, Germans don't like other people, you say, close your eyes and see and just see, then this means in other words that you can see a little bit of the acceptance of technology when making a decision and the most you say that inaction means that a comparison between Germany and France reveals cultural differences.

The French tend to favor respectful women and there is a greater focus on children and, unlike the Germans, the French do not want to leave things to fate. They want the machine to do things. decision the machine is a kind of mirror for the first time something that you did unconsciously or maybe instinctively in the case of an accident you know that you just act randomly now you have to make a conscious decision and the machine is forcing you to make a decision right, You can't, you can't wave it with your hand because in the end you have to program something.

Self-driving cars are not yet ready for the road, and ethical questions still abound. Artificial intelligence harbors immense potential to benefit daily life, medicine or mobility, but we must also look beyond the technical possibilities, what is that progress for? It is a question that artificial intelligence algorithms cannot answer, only humanity can.