Artificial Intelligence For Dummies book cover

Artificial Intelligence For Dummies

By: John Paul Mueller and Luca Massaron Published: 11-24-2021

Forget far-away dreams of the future. Artificial intelligence is here now! 

Every time you use a smart device or some sort of slick technology—be it a smartwatch, smart speaker, security alarm, or even customer service chat box—you’re engaging with artificial intelligence (AI). If you’re curious about how AI is developed—or question whether AI is real—Artificial Intelligence For Dummies holds the answers you’re looking for. Starting with a basic definition of AI and explanations of data use, algorithms, special hardware, and more, this reference simplifies this complex topic for anyone who wants to understand what operates the devices we can’t live without.  

This book will help you: 

  • Separate the reality of artificial intelligence from the hype 
  • Know what artificial intelligence can accomplish and what its limits are 
  • Understand how AI speeds up data gathering and analysis to help you make informed decisions more quickly 
  • See how AI is being used in hardware applications like drones, robots, and vehicles 
  • Know where AI could be used in space, medicine, and communication fields sooner than you think 

Almost 80 percent of the devices you interact with every day depend on some sort of AI. And although you don’t need to understand AI to operate your smart speaker or interact with a bot, you’ll feel a little smarter—dare we say more intelligent—when you know what’s going on behind the scenes.  So don’t wait. Pick up this popular guide to unlock the secrets of AI today! 

Articles From Artificial Intelligence For Dummies

page 1
page 2
page 3
page 4
35 results
35 results
AI For Dummies Cheat Sheet

Cheat Sheet / Updated 02-25-2022

Artificial Intelligence (AI) is a technology that has grabbed a lot of attention in movies, books, products, and in a slew of other places. Often, vendors equate AI with smartness: You buy a smart device to obtain a device with an AI, even though smart devices sometimes are smart only in that they offer connectivity, not AI. Many products are hyped to contain AI that sometimes doesn’t even work. Some people, of course, want to grab headlines by telling mistruths or offering misconceptions about. This cheat sheet doesn't up all the misconceptions, mistruths, and hype for you, but it does offer you some interesting insights into why the mundane uses of AI are actually the places where you see AI most often. Yes, AI is being put to some amazing uses as well, but vendors often misrepresent these uses to the point that no one really knows how much is real and how much is the result of someone’s vivid imagination.

View Cheat Sheet
Increase Hardware Capabilities for Artificial Intelligence

Article / Updated 11-14-2019

The CPU still works well for business systems or in applications in which the need for general flexibility in programming outweighs pure processing power. However, GPUs are now the standard for various kinds of data science, machine learning, AI, and deep-learning needs. Of course, everyone is constantly looking for the next big thing in the development environment. Both CPUs and GPUs are production-level processors. In the future, you may see one of two kinds of processors used in place of these standards: Application Specific Integrated Circuits (ASICs): In contrast to general processors, a vendor creates an ASIC for a specific purpose. An ASIC solution offers extremely fast performance using very little power, but it lacks flexibility. An example of an ASIC solution is Google’s Tensor Processing Unit (TPU), which is used for speech processing. Field Programmable Gate Arrays (FPGAs): As with an ASIC, a vendor generally crafts a FPGA for a specific purpose. However, contrary to an ASIC, you can program a FPGA to change its underlying functionality. An example of a FPGA solution is Microsoft’s Brainwave, which is used for deep-learning projects. The battle between ASICs and FPGAs promises to heat up, with AI developers emerging as the winner. For the time being, Microsoft and FPGAs appear to have taken the lead. The point is that technology is fluid, and you should expect to see new developments. Vendors are also working on entirely new processing types, which may or may not actually work as expected. For example, Graphcore is working on an Intelligence Processing Unit (IPU). You have to take the news of these new processors with a grain of salt given the hype that has surrounded the industry in the past. When you see real applications from large companies such as Google and Microsoft, you can start to feel a little more certain about the future of the technology involved.

View Article
New Surgical Techniques and Artificial Intelligence

Article / Updated 06-18-2019

Robots and artificial intelligence (AI) routinely participate in surgical procedures today. In fact, some surgeries would be nearly impossible without the use of robots and AI. However, the history of using this technology isn’t very lengthy. The first surgical robot, Arthrobot, made its appearance in 1983. Even so, the use of these life-saving technologies has reduced errors, improved results, decreased healing time, and generally made surgery less expensive over the long run. Making surgical suggestions You can view the whole idea of surgical suggestions in many different ways. For example, an AI could analyze all the data about a patient and provide the surgeon with suggestions about the best approaches to take based on that individual patient’s record. The surgeon could perform this task, but it would take longer and might be subject to errors that the AI won’t make. The AI doesn’t get tired or overlook things; it consistently views all the data available in the same way every time. Unfortunately, even with an AI assistant, surprises still happen during surgery, which is where the next level of suggestion comes into play. According to this article, doctors can now have access to a device that works along the same lines as Alexa, Siri, and Cortana (the AI in devices you may actually have in your own home). No, the device won’t take the doctor’s request for music to play during the surgery, but the surgeon can use the device to locate specific bits of information without having to stop. This means that the patient receives the benefit of what amounts to a second opinion to handle unforeseen complications during a surgery. Mind you, the device isn’t actually doing anything more than making already existing research created by other doctors readily available in response to surgeon requests; no real thinking is involved. Getting ready for surgery also means analyzing all those scans that doctors insist on having. Speed is an advantage that AI has over a radiologist. Products such as Enlitic, a deep-learning technology, can analyze radiological scans in milliseconds — up to 10,000 times faster than a radiologist. In addition, the system is 50 percent better at classifying tumors and has a lower false-negative rate (0 percent versus 7 percent) than humans. Another product in this category, Arterys, can perform a cardiac scan in 6 to 10 minutes, rather than the usual hour. Patients don’t have to spend time holding their breath, either. Amazingly, this system obtains several dimensions of data: 3-D heart anatomy, blood-flow rate, and blood-flow direction, in this short time. Watch this video about Arterys. Assisting a surgeon Most robotic help for surgeons today assists, rather than replaces, the surgeon. The first robot surgeon, the PUMA system, appeared in 1986. It performed an extremely delicate neurosurgical biopsy, which is a nonlaparoscopic type of surgery. Laparoscopic surgery is minimally invasive, with one or more small holes serving to provide access to an organ, such as a gall bladder, for removal or repair. The first robots weren’t adept enough to perform this task. By 2000, the da Vinci Surgical System provided the ability to perform robotic laparoscopic surgery using a 3-D optical system. The surgeon directs the robot’s movements, but the robot performs the actual surgery. The surgeon watches a high-definition display during the surgery and can actually see the operation better than being in the room performing the task personally. The da Vinci System also uses smaller holes than a surgeon can, reducing the risk of infection. The most important aspect of the da Vinci Surgical System, though, is that the setup augments the surgeon’s native capabilities. For example, if the surgeon shakes a bit during part of the process, the da Vinci Surgical System removes the shake — similarly to how anti-shake features work with a camera. The system also smoothes out external vibration. The system’s set up also enables the surgeon to perform extremely fine movements — finer than a human can natively perform, making the surgery far more precise than the surgeon could accomplish alone. The da Vinci Surgical System is a complex and extremely flexible device. The FDA has approved it for both pediatric and adult surgeries of the following types: Urological surgeries General laparoscopic surgeries General noncardiovascular thoracoscopic surgeries Thoracoscopically assisted cardiotomy procedures The point behind including all this medical jargon is that the da Vinci Surgical System can perform many tasks without involving a surgeon directly. At some point, robot surgeons will become more autonomous, keeping humans even further away from the patient during surgery. In the future, no one will actually enter the clean room with the patient, thereby reducing the chances of infection to nearly zero. You can read more about the da Vinci Surgical System. Replacing the surgeon with monitoring In Star Wars, you see robotic surgeons patching up humans all the time. In fact, you might wonder whether any human doctors are available. Theoretically, robots could take over some types of surgery in the future, but the possibility is still a long way off. Robots would need to advance quite a bit from the industrial sort of applications that you find today. The robots of today are hardly autonomous and require human intervention for setups. However, the art of surgery for robots is making advances. For example, the Smart Tissue Autonomous Robot (STAR) outperformed human surgeons when sewing a pig intestine. Doctors supervised STAR during the surgery, but the robot actually performed the task on its own, which is a huge step forward in robotic surgery. This video is quite informative about where surgery is going.

View Article
Performing Tasks Using Automation

Article / Updated 07-11-2018

Artificial intelligence (AI) is great at automation. It never deviates from the procedure, never gets tired, and never makes mistakes as long as the initial procedure is correct. Unlike humans, AI never needs a vacation or a break or even an eight-hour day (not that many in the medical profession have that, either). Consequently, the same AI that interacts with a patient for breakfast will do so for lunch and dinner as well. So, at the outset, AI has some significant advantages if viewed solely on the bases of consistency, accuracy, and longevity. Working with medical records The major way in which an AI helps in medicine is medical records. In the past, everyone used paper records to store patient data. Each patient might also have a blackboard that medical personnel use to record information daily during a hospital stay. Various charts contain patient data, and the doctor might also have notes. Having all these sources of information in so many different places made it hard to keep track of the patient in any significant way. Using an AI, along with a computer database, helps make information accessible, consistent, and reliable. Products such as Google Deepmind Health enable personnel to mine the patient information to see patterns in data that aren’t obvious. Doctors don’t necessarily interact with records in the same way that everyone else does. The use of products such as IBM’s WatsonPaths helps doctors interact with patient data of all sorts in new ways to make better diagnostic decisions about patient health. You can see a video on how this product works. Medicine is about a team approach, with many people of varying specialties working together. However, anyone who watches the process for a while soon realizes that these people don’t communicate among themselves sufficiently because they’re all quite busy treating patients. Products such as CloudMedX take all the input from the all parties involved and performs risk analysis on it. The result is that the software can help locate potentially problematic areas that could reduce the likelihood of a good patient outcome. In other words, this product does some of the talking that the various stakeholders would likely do if they weren’t submerged in patient care. Predicting the future Some truly amazing predictive software based on medical records includes CareSkore, which actually uses algorithms to determine the likelihood of a patient’s requiring readmission into the hospital after a stay. By performing this task, hospital staff can review reasons for potential readmission and address them before the patient leaves the hospital, making readmission less likely. Along with this strategy, Zephyr Health helps doctors evaluate various therapies and choose those most likely to result in a positive outcome — again reducing the risk that a patient will require readmission to the hospital. This video tells you more about Zephyr Health. In some respects, your genetics form a map of what will happen to you in the future. Consequently, knowing about your genetics can increase your understanding of your strengths and weaknesses, helping you to live a better life. Deep Genomics is discovering how mutations in your genetics affect you as a person. Mutations need not always produce a negative result; some mutations actually make people better, so knowing about mutations can be a positive experience, too. Check out this video for more details. Making procedures safer Doctors need lots of data to make good decisions. However, with data being spread out all over the place, doctors who lack the ability to analyze that disparate data quickly often make imperfect decisions. To make procedures safer, a doctor needs not only access to the data but also some means of organizing and analyzing it in a manner reflecting the doctor’s specialty. One such product is Oncora Medical, which collects and organizes medical records for radiation oncologists. As a result, is these doctors can deliver the right amount of radiation to just the right locations to obtain a better result with a lower potential for unanticipated side effects. Doctors also have trouble obtaining necessary information because the machines they use tend to be expensive and huge. An innovator named Jonathan Rothberg has decided to change all that by using the Butterfly Network. Imagine an iPhone-sized device that can perform both an MRI and an ultrasound. The picture on the website is nothing short of amazing. Creating better medications Everyone complains about the price of medications today. Yes, medications can do amazing things for people, but they cost so much that some people end up mortgaging homes to obtain them. Part of the problem is that testing takes a lot of time. Performing a tissue analysis to observe the effects of a new drug can take up to a year. Fortunately, products such as 3Scan can greatly reduce the time required to obtain the same tissue analysis to as little as one day. Of course, better still would be the drug company having a better idea of which drugs are likely to work and which aren’t before investing any money in research. Atomwise uses a huge database of molecular structures to perform analyses on which molecules will answer a particular need. In 2015, researchers used Atomwise to create medications that would make Ebola less likely to infect others. The analysis that would have taken human researchers months or possibly years to perform took Atomwise just one day to complete. Imagine this scenario in the midst of a potentially global epidemic. If Atomwise can perform the analysis required to render the virus or bacteria noncontagious in one day, the potential epidemic could be curtailed before becoming widespread. Drug companies also produce a huge number of drugs. The reason for this impressive productivity, besides profitability, is that every person is just a little different. A drug that performs well and produces no side effects on one person might not perform well at all and could even harm a different person. Turbine enables drug companies to perform drug simulations so that the drug companies can locate the drugs most likely to work with a particular person’s body. Turbine’s current emphasis is on cancer treatments, but it’s easy to see how this same approach could work in many other areas. Medications can take many forms. Some people think they come only in pill or shot form, yet your body produces a wide range of medications in the form of microbiomes. Your body actually contains ten times as many microbes as it does human cells, and many of these microbes are essential for life; you’d quickly die without them. Whole Biome is using a variety of methods to make these microbiomes work better for you so that you don’t necessarily need a pill or a shot to cure something. Check out this video for additional information. Some companies have yet to realize their potential, but they’re likely to do so eventually. One such company is Recursion Pharmaceuticals, which employs automation to explore ways to use known drugs, bioactive drugs, and pharmaceuticals that didn’t previously make the grade to solve new problems. The company has had some success in helping to solve rare genetic diseases, and it has a goal of curing 100 diseases in the next ten years (obviously, an extremely high goal to reach).

View Article
Artificial Intelligence and Special Needs

Article / Updated 07-11-2018

At one time, losing a limb or having another special need meant years of doctor visits, reduced capability, and a shorter and less happy life. However, better prosthetics and other devices, many of them AI-enabled, have made this scenario a thing of the past for many people. For example, check out this couple dancing. The woman has a prosthetic leg. These days, some people can run a marathon or go rock climbing, even if they’ve lost their original legs. Many people view the term special needs as being equivalent to physically or mentally deficient or even disabled. However, just about everyone has some special need. At the end of a long day, someone with perfectly normal vision might benefit from magnifying software to make text or graphic elements larger. Color translation software can help someone with normal color vision see details that aren’t normally visible (at least, to someone with what is considered normal vision). As people get older, they tend to need more assistance to hear, see, touch, or otherwise interact with common objects. Likewise, assistance with tasks such as walking could keep someone out of a nursing home and in their own home for their entire life. The point is that using various kinds of AI-enabled technologies can significantly help everyone to have a better life. Considering the software-based solutions Many people using computers today rely on some type of software-based solution to meet specific needs. One of the most famous of these solutions is a screen reader called Job Access With Speech (JAWS) that tells you about display content using sophisticated methods. As you might imagine, every technique that both data science and AI rely upon to condition data, interpret it, and then provide a result likely occurs within the JAWS software, making it a good way for anyone to understand the capabilities and limits of software-based solutions. The best way for you to see how this works for you is to download and install the software, and then use it while blindfolded to perform specific tasks on your system. (Avoid anything that will terrify you, though, because you’ll make mistakes.) Accessibility software helps people with special needs perform incredible tasks. It can also help others understand what it would be like to have a special need. A considerable number of such applications are available, but check out Vischeck at or one example. This lets you see graphics in the same way that people with specific kinds of color blindness see them. Of course, the first thing you’ll discover is that the term color blind is actually incorrect; people with these conditions see color just fine. The color is simply shifted to a different color, so saying color shifted is likely a better term. Relying on hardware augmentation Many kinds of special needs require more than just software to address adequately. The “Considering the use of exoskeletons” section, earlier in this chapter, tells you about the various ways in which exoskeletons see use today in preventing injury, augmenting natural human capabilities, or addressing special needs (such as allowing a paraplegic to walk). However, many other kinds of hardware augmentation address other needs, and the vast majority require some level of AI to work properly. Consider, for example, the use of eye-gaze systems. The early systems relied on a template mounted on top of the monitor. A quadriplegic could look at individual letters, which would be picked up by two cameras (one on each side of the monitor) and then typed into the computer. By typing commands this way, the quadriplegic could perform basic tasks at the computer. Some of the early eye-gaze systems connected to a robotic arm through the computer. The robotic arm could do extremely simple but important actions, such as help users get a drink or scratch their nose. Modern systems actually help connect a user’s brain directly to the robotic arm, making it possible to perform tasks such as eating without help. Seeing AI in prosthetics You can find many examples of AI used in prosthetics. Yes, some passive examples exist, but most of the newer visions for prosthetics rely on dynamic approaches that require an AI to perform. One of the more amazing examples of AI-enabled prosthetics is the fully dynamic foot created by Hugh Herr. This foot and ankle work so well that it’s actually possible for Hugh to perform tasks such as rock climbing. You can see a presentation that he made recently at TED. A moral dilemma that we might have to consider sometime in the future (thankfully not today) is when prosthetics actually allow their wearers to substantially surpass native human capability. For example, in the movie Eon Flux, Sithandra has hands for feet. The hands are essentially a kind of prosthetic grafted to someone who used to have normal feet. The question arises as to whether this kind of prosthetic implementation is valid, useful, or even desirable. At some point, a group of people will need to sit down and ascertain where prosthetic use should end to maintain humans as humans (assuming that we decide to remain human and not evolve into some next phase). Obviously, you won’t see anyone with hands for feet today.

View Article
Artificial Intelligence and Making Humans More Capable

Article / Updated 07-11-2018

Many of the current techniques for extending the healthy range of human life (the segment of life that contains no significant sickness), rather than just increasing the number of years of life depends on making humans more capable of improving their own health in various ways. You can find any number of articles that tell you 30, 40, or even 50 ways to extend this healthy range, but often it comes down to a combination of eating right, exercising enough and in the right way, and sleeping well. Of course, figuring out just which food, exercise, and sleep technique works best for you is nearly impossible. The following sections discuss ways in which an AI-enabled device might make the difference between having 60 good years and 80 or more good years. (In fact, it’s no longer hard to find articles that discuss human life spans of 1,000 or more years in the future because of technological changes.) Using games for therapy A gaming console can make a powerful and fun physical therapy tool. Both Nintendo Wii and Xbox 360 see use in many different physical therapy venues. The goal of these games is to get people moving in certain ways. As when anyone else plays, the game automatically rewards proper patient movements, but a patient also receives therapy in a fun way. Because the therapy becomes fun, the patient is more likely to actually do it and get better faster. Of course, movement alone, even when working with the proper game, doesn’t assure success. In fact, someone could develop a new injury when playing these games. The Jintronix add-on for the Xbox Kinect hardware standardizes the use of this game console for therapy, increasing the probability of a great outcome. Considering the use of exoskeletons One of the most complex undertakings for an AI is to provide support for an entire human body. That’s what happens when someone wears an exoskeleton (essentially a wearable robot). An AI senses movements (or need to move) and provides a powered response to the need. The military has excelled in the use of exoskeletons. Imagine being able to run faster and carry significantly heavier loads as a result of wearing an exoskeleton. This video gives you just a glimpse of what’s possible. Of course, the military continues to experiment, which actually feeds into civilian uses. The exoskeleton you eventually see (and you’re almost guaranteed to see one at some point) will likely have its origins in the military. Industry has also gotten in on the exoskeleton technology. Factory workers currently face a host of illnesses because of repetitive stress injuries. In addition, factory work is incredibly tiring. Wearing an exoskeleton not only reduces fatigue but also reduces errors and makes the workers more efficient. People who maintain their energy levels throughout the day can do more with far less chance of being injured, damaging products, or hurting someone else. The exoskeletons in use in industry today reflect their military beginnings. Look for the capabilities and appearance of these devices to change in the future to look more like the exoskeletons shown in movies such as Aliens. The real-world examples of this technology are a little less impressive but will continue to gain in functionality. As interesting as the use of exoskeletons to make able people even more incredible is, what they can enable people to do that they can’t do now is downright amazing. For example, a recently published Smithsonian article discusses using an exoskeleton to enable a child with cerebral palsy to walk. Not all exoskeletons used in medical applications provide lifetime use, however. For example, an exoskeleton can help a stroke victim walk normally again. As the person becomes more able, the exoskeleton provides less support until the wearer no longer needs it. Some users of the device have even coupled their exoskeleton to other products, such as Amazon’s Alexa. The overall purpose of wearing an exoskeleton isn’t to make you into Iron Man. Rather, it’s to cut down on repetitive stress injuries and help humans excel at tasks that currently prove too tiring or just beyond the limits of their body. From a medical perspective, using an exoskeleton is a win because it keeps people mobile longer, and mobility is essential to good health.

View Article
Portable Patient Monitoring

Article / Updated 07-11-2018

A medical professional isn’t always able to tell what is happening with a patient’s health simply by listening to their heart, checking vitals, or performing a blood test. The body doesn’t always send out useful signals that let a medical professional learn anything at all. In addition, some body functions, such as blood sugar, change over time, so constant monitoring becomes necessary. Going to the doctor’s office every time you need one of these vitals checked would prove time consuming and possibly not all that useful. Older methods of determining some body characteristics required manual, external intervention on the part of the patient — an error-prone process in the best of times. For these reasons, and many more, an AI can help monitor a patient’s statistics in a manner that is efficient, less error prone, and more consistent, as described in the following sections. Wearing helpful monitors All sorts of monitors fall into the helpful category. In fact, many of these monitors have nothing to do with the medical profession, yet produce positive results for your health. Consider the Moov monitor, which monitors both heart rate and 3-D movement. The AI for this device tracks these statistics and provides advice on how to create a better workout. You actually get advice on, for example, how your feet are hitting the pavement during running and whether you need to lengthen your stride. The point of devices like these is to ensure that you get the sort of workout that will improve health without risking injury. Mind you, if a watch-type monitoring device is too large, Motiv produces a ring that monitors about the same number of things that Moov does, but in a smaller package. This ring even tracks how you sleep to help you get a good night’s rest. Rings do tend to come with an assortment of pros and cons. This article tells you more about these issues. Interestingly enough, many of the pictures on the site don’t look anything like a fitness monitor, so you can have fashion and health all in one package. Of course, if your only goal is to monitor your heart rate, you can get devices such as Apple Watch that also provide some level of analysis using an AI. All these devices interact with your smartphone, so you can possibly link the data to still other applications or send it to your doctor as needed. Relying on critical wearable monitors A problem with some human conditions is that they change constantly, so checking intermittently doesn’t really get the job done. Glucose, the statistic measured by diabetics, is one statistic that falls into this category. The more you monitor the rise and fall of glucose each day, the easier it becomes to change medications and lifestyle to keep diabetes under control. Devices such as the K'Watch provide such constant monitoring, along with an app that a person can use to obtain helpful information on managing their diabetes. Of course, people have used intermittent monitoring for years; this device simply provides that extra level of monitoring that can make the difference between having diabetes be a life-altering issue or a minor nuisance. The act of constantly monitoring someone’s blood sugar or other chronic disease statistic might seem like overkill, but it has practical use as well. Products such as Sentrian let people use the remote data to predict that a patient will become ill before the event actually occurs. By making changes in patient medications and behavior before an event can occur, Sentrian reduces the number of unavoidable hospitalizations — making the patient’s life a lot better and reducing medical costs. Some devices are truly critical, such as the Wearable Defibrillator Vest (WDV), which senses your heart condition continuously and provides a shock should your heart stop working properly. This short-term solution can help a doctor decide whether you need the implanted version of the same device. There are pros and cons to wearing one, but then again, it’s hard to place a value on having a shock available when needed to save a life. The biggest value of this device is the monitoring it provides. Some people don’t actually need an implantable device, so monitoring is essential to prevent unnecessary surgery. Using movable monitors The number and variety of AI-enabled health monitors on the market today is staggering. For example, you can actually buy an AI-enabled toothbrush that will monitor your brushing habits and provide you with advice on better brushing technique. When you think about it, creating a device like this presents a number of hurdles, not the least of which is keeping the monitoring circuitry happy inside the human mouth. Of course, some people may feel that the act of brushing their teeth really doesn’t have much to do with good health, but it does. Creating movable monitors generally means making them both smaller and less intrusive. Simplicity is also a requirement for devices designed for use by people with little or no medical knowledge. One device in this category is a wearable electrocardiogram (ECG). Having an ECG in a doctor’s office means connecting wires from the patient to a semiportable device that performs the required monitoring. The QardioCore provides the ECG without using wires, and someone with limited medical knowledge can easily use it. As with many devices, this one relies on your smartphone to provide needed analysis and make connections to outside sources as needed. Current medical devices work just fine, but they aren’t portable. The point of creating AI-enabled apps and specialized devices is to obtain much needed data when a doctor actually needs it, rather than having to wait for that data. Even if you don’t buy a toothbrush to monitor your technique or an ECG to monitor your heart, the fact that these devices are small, capable, and easy to use means that you may still benefit from them at some point.

View Article
Artificial Intelligence and Safe Environments

Article / Updated 07-11-2018

One of the most often stated roles for AI, besides automating tasks, is keeping humans safe in various ways. Articles such as this one describe an environment in which AI acts as an intermediary, taking the hit that humans would normally take when a safety issue occurs. Safety takes all sorts of forms. Yes, AI will make working in various environments safer, but it’ll also help create a healthier environment and reduce risks associated with common tasks, including surfing the Internet. The role of boredom in accidents From driving to work, boredom increases accidents of all sorts. In fact, anytime that someone is supposed to perform a task that requires any level of focus and instead acts in a somnolent manner, the outcome is seldom good. The problem is so serious and significant that you can find a wealth of articles on the topic, such as “Modelling human boredom at work: mathematical formulations and a probabilistic framework.” Whether an accident actually occurs (or was a close call) depends on random chance. Imagine actually developing algorithms that help determine the probability of accidents happening due to boredom under certain conditions. AI in avoiding safety issues No AI can prevent accidents owing to human causes, such as boredom. In a best-case scenario, when humans decide to actually follow the rules that AI helps create, the AI can only help avoid potential problems. Unlike with Asimov’s robots, there are no three-laws protections in place in any environment; humans must choose to remain safe. With this in mind, an AI could help in these ways: Suggest job rotations (whether in the workplace, in a car, or even at home) to keep tasks interesting Monitor human performance in order to better suggest down time because of fatigue or other factors Assist humans in performing tasks to combine the intelligence that humans provide with the quick reaction time of the AI Augment human detection capabilities so that potential safety issues become more obvious Take over repetitive tasks so that humans are less likely to become fatigued and participate in the interesting aspects of any job AI can’t eliminate safety issues Ensuring complete safety implies an ability to see the future. Because the future is unknown, the potential risks to humans at any given time are also unknown because unexpected situations can occur. An unexpected situation is one that the original developers of a particular safety strategy didn’t envision. Humans are adept at finding new ways to get into predicaments, partly because we’re both curious and creative. Finding a method to overcome the safety provided by an AI is in human nature because humans are inquisitive; we want to see what will happen if we try something — generally something stupid. Unpredictable situations aren’t the only problem that an AI faces. Even if someone were to find every possible way in which a human could become unsafe, the processing power required to detect the event and determine a course of action would be astronomical. The AI would work so slowly that its response would always occur too late to make any difference. Consequently, developers of safety equipment that actually requires an AI to perform the required level of safety have to deal in probabilities and then protect against the situations that are most likely to happen.

View Article
Artificial Intelligence Solutions for Boredom

Article / Updated 07-11-2018

Polls often show what people think they want, rather than what they do want, but they’re still useful. When polled to see what kind of life recent college graduates wanted, not one of them said boredom. In fact, you could possibly poll just about any group and not come up with a single boring response. Most humans (saying all would likely result in an avalanche of email with examples) don’t want to be bored. In some cases, AI can work with humans to make life more interesting — for the human, at least. Making tasks more interesting Any occupation, be it personal or for an organization, has certain characteristics that attract people and makes them want to participate in it. Obviously, some occupations, such as taking care of your own children, pay nothing, but the satisfaction of doing so can be incredibly high. Likewise, working as a bookkeeper may pay quite well but not offer much in the way of job satisfaction. Various polls and articles talk about the balance of money and satisfaction, but reading them often proves confusing because the basis for making a determination is ambiguous. However, most of these sources agree that after a human makes a certain amount of money, satisfaction becomes the key to maintaining interest in the occupation (no matter what that occupation might be). Of course, figuring out what comprises job satisfaction is nearly impossible, but interest remains high on the list. An interesting occupation will always have higher satisfaction potential. The problem is not one of necessarily changing jobs, then, but of making the job more interesting as a means to avoid boredom. An AI can effectively help this process by removing repetition from tasks. However, examples such as Amazon’s Alexa and Google’s Home do provide other alternatives. The feeling of loneliness that can pervade the home, workplace, car, and other locations is a strong creator of boredom. When humans begin to feel alone, depression sets in and boredom is often just a step away. Creating applications that use the Alexa interface or Actions on Google API to simulate human interaction of the appropriate sort can improve the workplace experience. More important, developing smart interfaces of this sort can help humans perform a wealth of mundane tasks quickly, such as researching information and interacting smart devices, not just light switches. Helping humans work more efficiently Most humans, at least the forward-thinking ones, have some ideas of how they’d like an AI to make their lives better by eliminating tasks that they don’t want to do. A recent poll shows some of the more interesting ways that AI can do this. Many of them are mundane, but notice the ones like detecting when a spouse is unhappy and sending flowers. It probably won’t work, but it’s an interesting idea nonetheless. The point is that humans will likely provide the most interesting ideas on how to create an AI that specifically addresses that person’s needs. In most cases, serious ideas will work well for other users, too. For example, automating trouble tickets is something that could work in a number of different industries. If someone were to come up with a generic interface, with a programmable back end to generate the required custom trouble tickets, the AI could save users a lot of time and ensure future efficiencies by ensuring that trouble tickets consistently record the required information. How AI reduces boredom Boredom comes in many packages, and humans view these packages in different ways. There is the boredom that comes from not having required resources, knowledge, or other needs met. Another kind of boredom comes from not knowing what to do next. An AI can help with the first kind of boredom; it can’t help with the second. This section considers the first kind of boredom. Access to resources of all sorts helps reduce boredom by allowing humans to create without the mundane necessity of acquiring needed materials. Here are some ways in which an AI can make access to resources easier: Searching for needed items online Ordering needed items automatically Performing sensor and other data-acquisition monitoring Managing data Accomplishing mundane or repetitive tasks How AI can’t reduce boredom An AI is not creative or intuitive. So, asking an AI to think of something for you to do is unlikely to produce satisfying results. Someone could program the AI to track the top ten things you like to do and then select one of them at random, but the result still won’t be satisfying because the AI can’t take aspects like your current state of mind into account. In fact, even with the best facial expression, an AI will lack the capability to interact with you in a manner that will produce any sort of satisfying result. An AI also can’t motivate you. Think about what happens when a friend comes by to help motivate you (or you motivate the friend). The friend actually relies on a combination of intrapersonal knowledge (empathizing by considering how she’d feel in your situation) and interpersonal knowledge (projecting creative ideas on how to obtain a positive emotional response from you). An AI won’t have any of the first kind of knowledge and only extremely limited amounts of the second kind of knowledge. Consequently, an AI can’t reduce your boredom through motivational techniques. Boredom may not always be a bad thing, anyway. A number of recent studies have shown that boredom actually helps promote creative thought, which is the direction that humans need to go. After viewing myriad articles on how AI is going to take jobs away, it’s important to consider that the jobs that AI is taking are, in themselves, often boring and leave humans no time to create. Even today, humans could find productive, creative, jobs to do if they really thought about it. The article “7 Surprising Facts About Creativity, According To Science” actually discusses the role of boring tasks like daydreaming in enhancing creativity. In the future, if humans really want to reach for the stars and do other amazing things, creativity will be essential, so the fact that AI can’t reduce your boredom is actually a good thing.

View Article
Making Suggestions and Artificial Intelligence

Article / Updated 07-11-2018

A suggestion is different from a command. Even though some humans seem to miss the point entirely, a suggestion is simply an idea put forth as a potential solution to a problem. Making a suggestion implies that other solutions could exist and that accepting a suggestion doesn’t mean automatically implementing it. In fact, the suggestion is only an idea; it may not even work. Of course, in a perfect world, all suggestions would be good suggestions — at least possible solutions to a correct output, which is seldom the case in the real world. Getting suggestions based on past actions The most common way that an AI uses to create a suggestion is by collecting past actions as events and then using those past actions as a dataset for making new suggestions. For example, someone purchases a Half-Baked Widget every month for three months. It makes sense to suggest buying another one at the beginning of the fourth month. In fact, a truly smart AI might make the suggestion at the right time of the month. For example, if the user makes the purchase between the third and the fifth day of the month for the first three months, it pays to start making the suggestion on the third day of the month and then move onto something else after the fifth day. Humans output an enormous number of clues while performing tasks. Unlike humans, an AI actually pays attention to every one of these clues and can record them in a consistent manner. The consistent collection of action data makes enables an AI to provide suggestions based on past actions with a high degree of accuracy in many cases. Getting suggestions based on groups Another common way to make suggestions relies on group membership. In this case, group membership need not be formal. A group could consist of a loose association of people who have some minor need or activity in common. For example, a lumberjack, a store owner, and a dietician could all buy mystery books. Even though they have nothing else in common, not even location, the fact that all three like mysteries makes them part of a group. An AI can easily spot patterns like this that might elude humans, so it can make good buying suggestions based on these rather loose group affiliations. Groups can include ethereal connections that are temporary at best. For example, all the people who flew flight 1982 out of Houston on a certain day could form a group. Again, no connection whatsoever exists between these people except that they appeared on a specific flight. However, by knowing this information, an AI could perform additional filtering to locate people within the flight who like mysteries. The point is that an AI can provide good suggestions based on group affiliation even when the group is difficult (if not impossible) to identify from a human perspective. Obtaining the wrong suggestions Anyone who has spent time shopping online knows that websites often provide suggestions based on various criteria, such as previous purchases. Unfortunately, these suggestions are often wrong because the underlying AI lacks understanding. When someone makes a once-in-a-lifetime purchase of a Super-Wide Widget, a human would likely know that the purchase is indeed once in a lifetime because it’s extremely unlikely that anyone will need two. However, the AI doesn’t understand this fact. So, unless a programmer specifically creates a rule specifying that Super-Wide Widgets are a once-in-a-lifetime purchase, the AI may choose to keep recommending the product because sales are understandably small. In following a secondary rule about promoting products with slower sales, the AI behaves according to the characteristics that the developer provided for it, but the suggestions it makes are outright wrong. Besides rule-based or logic errors in AIs, suggestions can become corrupted through data issues. For example, a GPS could make a suggestion based on the best possible data for a particular trip. However, road construction might make the suggested path untenable because the road is closed. Of course, many GPS applications do consider road construction, but they sometimes don’t consider other issues, such as a sudden change in the speed limit or weather conditions that make a particular path treacherous. Humans can overcome lacks in data through innovation, such as by using the less traveled road or understanding the meaning of detour signs. When an AI manages to get past the logic, rule, and data issues, it sometimes still makes bad suggestions because it doesn’t understand the correlation between certain datasets in the same way a human does. For example, the AI may not know to suggest paint after a human purchases a combination of pipe and drywall when making a plumbing repair. The need to paint the drywall and the surrounding area after the repair is obvious to a human because a human has a sense of aesthetics that the AI lacks. The human makes a correlation between various products that isn’t obvious to the AI.

View Article
page 1
page 2
page 3
page 4