AI and data science has Innumerable Applications. Let’s look at 100 use-cases of AI, ML and data science.
We, as humans, are shaped by our experiences. Touch a hot stove element as a child, and you learn quickly never to do it again. Spin and spin with your friends on the front lawn until you fall down, to experience the feeling that they sky is spinning around you. Keep doing it over and over until you throw up, and you have learned a limitation. You may do it again, but you’ve learned to stop doing it sooner because of the consequences.
Artificial Intelligence (AI) does not possess the amazing feedback system that has evolved on this planet for everything with a brain and sophisticated nervous system weighing more than a few grams. Instead, we have to define what constitutes success or failure for our increasingly clever AIs, and they need to assess everything they learn through a “Is this what humans would want?” filter.
They fit in every industry, whether manufacturing or service, and so there is not much point in specifying how we will use it in “Real Estate” or “Biotechnology”. Instead there will be many instantiations that fit multiple industries.
Collecting data encompassing the accumulated wisdom of thousands of experts, an AI System could access these details to address any problem that arose on that subject. Combined Expert Systems will have a vast interdisciplinary knowledge capable of solving some of humanity’s toughest questions.
Drug Discovery Research Development
During an Ebola outbreak, instead of the typical two to three years of development time, one Pharma company set its AI loose on its drug molecular database and within a day it had come up with two candidates suitable for Clinical Trials. AIs discover relationships that humans beings miss, which means they will aid us in discovering essential new drugs (such as a new antibiotic, which we desperately need) or drug combination therapies to solve unique problems that arise.
Online AI Physician Advice
Britain and other countries already have AI Physician Expert Systems available to the public for a cost of about £60 per year. You can check symptoms anytime, 24 hours per day, and get practical medical advice as often as you wish, without limitation. This frees physicians to deal with people in person and have time to address their needs more thoroughly.
All markets are in need of interpretation for commercial purposes. The one that usually comes to mind first is the stock market. An AI with access to relevant news can predict stock prices, however, in day-to-day business it can predict customer’s evolving desires so that you have enough bathing suits, snow shovels, or pimento stuffed olives on hand, too.
AI Traders are like human Day Traders on steroids. A human takes a few moments to watch a stock, guess where it is going, buy into it, (hopefully) see the profit, and then to quickly sell it. Many Day Traders will only stay in a stock for a few seconds before moving on to their next purchase. An AI has a constant stream of data, understanding how it changes relative to a dozen or a hundred other changes. It can buy and sell numerous stocks concurrently, with absolute focus on the Big Picture, something no human can replicate.
Early Childhood Education
An AI could be a clever companion for a child. It could detect the emotional state, or intellectual focus, and provide content to match the current environment. This could be especially useful for autistic or other special needs children who respond better to non-human caregivers that could be in the form of pets/animals that they like.
Robotic Control CAD/CAM
Computer Aided Design and Computer Aided Manufacturing allow the creation of functional or decorative objects. If a customer wanted a dollhouse for a child’s gift, and wanted it to resemble their actual home, they could submit external pictures from all angles, and inform the AI that “This roofline is 36 feet long. This roof peak is 24 feet high,” and the AI could extrapolate all the other sizes. They could specify that the finished model should be three feet tall, and the AI could define plans for a 3D printer to create the object. This will be extremely common in the near future.
Instead of programming each step in a process manually, Robotic Process Automation simple observes a human accomplish a task and then replicates the effort. The difference between something drive by PLCs (Programmable Logic Controllers) and one drive by AI, is that the AI can learn as it works, and cope with atypical situations. If a part arrives upside down at an assembly station, the AI can recognize the problem, correct it, and carry on.
Right now we’re building autonomous marine investigators that explore the sea. They monitor temperatures, salinity, pollution, wave action, seismic or volcanic activity, and can stay submerged for months. When their duty cycle is complete, they simply deploy a bag filled with gas and rise to the surface, transmitting a “Pick me up!” signal. NASA preforms remote sensing tasks with planetary probes and robots that follow instructions, but autonomous in deciding the best way to accomplish the task when they are light-seconds, minutes, or even light-hours away.
Smartphones with built-in AI (Digital ME)
By 2020, it is estimated that all new Smartphones will be equipped with at least a modest AI that learns your own behaviors and can even “represent” you when you are unable to respond. The longer it works with you, the better it will be doing whatever “you” tasks you assign to it. Better than Siri or Alexa, this will actually understand you, and learn your habits so it can prepare information you’ll need just as you need it.
Snore-Detection with Wristband
AIs either in a home system, or located of a smart device, can manage Active Listening tasks. If you snore, it can buzz your smartwatch to simulate you to shift positions without actually waking you up. Tales of “I’ve fallen and I can’t get up” will not require someone to push a button and communicate. The AI can detect status changes, such as heart rate, breathing, or lack of movement and call appropriate support services.
Computer Vision for interpreting Imagery
A ChatBot can carry on a very personalized conversation now, and they can isolate goods that a customer wants by understanding “red cardigan sweater, cable-knit, with a white plaited lace collar” and find what the customer wants from inventory. Computer Vision will allow much better correlation when a customer can drop an image of the item in the chat.
Computer Image Recognition for Customer Service
AI is going to take over a lot of claims settlement for insurance companies. It is going to have to interpret images of houses, cars, fences, sheds, boats or whatever else might be insured, such as fields of produce, or livestock. It knows what something should look like, and it will need to be instructed on assessing damage. Insurance companies, dealing with major floods or storms will send out assessor drones to cover vast tracts of land quickly. No humans are necessary, and it is much faster. Their imagery can also assist Aid Workers on site.
Multisource Big Data Integration for Development
Collecting data from many disparate sources is part of the process of building “Big Data” repositories. It can include documentation, of course, but it can also include video data, voice recordings, and photography. AI will sort and classify this vast pool of data into a useable form.
Google Glass was too much, too soon. It was perfectly obvious that it was in use, and the sense of distraction was palpable. Newer versions, much more subtle, can layer imagery over your everyday environment, showing the location of items you are shopping for, with information such as cost, nutrition, or if another store has it for less money. Being AI it will understand that you’re likely to ask for the weather at a certain time, or want reminders about meetings so it will simply “pop up” unobtrusively.
App Development on Demand for single users
Programming takes a certain amount of creativity, the realm of humans. For Apps that need to perform a specific function, an AI with an Object-Oriented toolbox can churn out a customized app in just a few seconds. Tell it what inputs you’ll provide, the outputs desired, and the app should be ready in seconds.
Meeting Participant — Listens and contributes Data in real time
An AI team-member listens to you as your team brainstorms ideas. Someone posits an idea and the AI calls up images of whatever is necessary to understand the point being made. With AR it could draw such an image in mid-air for everyone to study.
Predictive System Maintenance to Minimize cost/downtime
This one is self-explanatory. There may never be a good time to take a machine out of production but an AI planning all of the production can arrange things to have the minimum impact.
Medical Imaging Analysis — Disease/Diagnostics
Nuclear, X-rays, MRIs, CAT, PET, and CT scans are all made better with an AI examining images. It can see detail at a level that eludes humans, finding relationships pointing to pre-cancerous lung tissue months before humans would spot it, or counting instances of micro-porosity in finger bones to determine someone’s age.
Medical Diagnostics — Heart Sound/Computed Tomography
One day you’ll be able to step into a small room and an AI Expert System physician program will be able to monitor your health while you sit in a special chair. The British AI Doctor program relies on patient descriptions to arrive at conclusions. Soon it will have empirical data to add into the mix for more accurate results.
Medical Treatment Plan Design (Pregnancy Support, et al)
Once a patient is diagnosed, that is only the beginning. Some patients respond to one therapy that is completely ineffective for another. An AI has the ability to look at all the patient details, including heredity, genotype, phenotype, history and create a plan that is effective for each slice of the population.
Stock Maintenance by Video in Retail Space
Most retail shops have video monitoring. It is a very small step to evolve to the point where active cameras track inventory at all times and place orders that minimize the amount in the warehouse, and maximize cash flow.
Shelf Space Analysis for Optimal Placement for Sales
As long as it is monitoring all the time, another useful ability is to analyze where something ought to be placed to maximize sales. Items that encourage the sales of other products should be placed closer together. Even items placed at certain heights can change sales significantly.
Emotion or Sentiment Analysis — Sales
AIs can extrapolate how people “feel” about a product, service, company, or event by assessing mood in small groups, or across the entire World Wide Web. You’ll know just as it begins if your company is becoming a thought-leader. With this warning, you’ll be able to exploit your popularity much more quickly.
Emotion or Sentiment Analysis — Advertising
AIs can also create individual advertising projects or entire campaigns designed to maximize the effect of public sentiment. Did you make a big public donation and now everyone loves you? A properly trained AI can keep the momentum going with PR releases, and sales changes.
Customer Clickable Videos when Products are Visible
You are going to be selling a lot more by video, whether it’s a tangible product or service. The key thing to remember is that almost everything is going to be touch screen compatible. AIs can invisibly tag everything seen on the screen. It could be items in the shelving unit, or it could be a particular arrangement of flowers in a shop. People inspired by images are ready to make a spontaneous purchase, so be ready for them.
Sorting your collected data is one of the most important advantages of AI. It can comb through your Big Data and reveal insights that would otherwise be missed.
There’s a difference between collecting and hoarding data compared to actually putting it into a useful form. It accumulates too quickly, with overwhelming quantities. AI will be the only way to interpret it all.
Connect Support Calls to best agent (using Caller ID and History)
When an identifiable repeat customer calls, an AI call router can direct that customer to an agent with whom they have a good relationship, or failing that, to a particular class of agent who deals with that sort of customer well.
An AI that understands our financial state can keep our bills paid in the most efficient manner, put aside money for desired acquisitions, and make sure that we attend to our investment and savings needs. It can warn us when we start to go astray.
Social Media Optimization for Businesses
JetBlue (among others) maintains a strong social media monitoring presence. If someone tweets a complaint about their company, their goal is to respond within 15 minutes. Meal coupons for delays, hotels for long delays, free tickets… whatever it takes. It is extensive, beyond the capability of a single human, but well-within the capability of an AI.
Physical Personal Assistants (Everyday/Common/Elderly)
The ageing population of Japan now outnumbers the younger generation and the rest of the world will soon follow. Care workers were being injured lifting the elderly during the move from one location to another (Bed > wheelchair > bathtub, etc.) Actual big, friendly looking white robots with stylized Panda faces are in use now, and can pick someone up from the floor, and put them in a chair, tub, or bed. Currently still overseen by a human, soon they can be autonomous, handling requests in spoken language.
Virtual Personal Assistants
Of course, we will continue to have Alexa, Lyra, Siri, G-Now, and so on, but they will be much, much better. Newer versions, with greater capabilities, will spawn like dandelions. These “Weak” AIs will become True AIs.
Heavy Industry Accident Prevention
Analyzing ongoing visual data, AIs can see emerging trends in workplace activity and predict safety issues before they manifest as a problem. Is the placement of equipment making employees alter their behavior, potentially conflicting with vehicle traffic? Are dosimeters not being returned at nuclear facilities? Eventually moving equipment will have no drivers, instead overseen by the AI, and there will never be an accident again.
An AI that understands our proclivities and tendencies, as well as our financial history, is in a good position to advise us on investments. It will know our preferences and risk tolerance, so whatever it comes up with should be ideally suited to us, as individuals. It can be a completely hands-off affair, where we just watch our retirement fund grow.
Financial Underwriting (Millennials with no credit score/history)
How do you get a loan if you need one, but have been living at home until age 33? Millennials have difficulty because they have no conventional credit history to study, so their credit history is almost blank. AIs can assess them based upon even a meager history, and predict what sort of risk they actually represent.
The AIs will find “true” risk and set rates accordingly. Thirty years with no citations, or accidents? $400 per year! One DWI, two speeding tickets while in a school zone, and a charge of Stunt Driving? $1,000 per month!
Actuarial tables are all about averages, zones, typicality, and biases. AI systems can rate each client individually based on the risk they represent. Someone who street races regularly yet has never been caught will inevitably brag on social media or elsewhere. The AI will know, or find out.
Massive Staff Scheduling
Having 20,000 employees is a challenge. Scheduling them, getting all the right people, with all the right skills, all in the right place… That a job for an AI!
Training in VR environment
Whether it is learning (or practicing) surgery, aircraft flight control, or lightsaber combat, AI controlled VR will be the training method of choice in the future.
Fitness Assessment (Drivers/Pilots)
Until we do away with flawed humans controlling planes, trains, trucks, and automobiles those people need to be recertified periodically. This can be in situ testing, built into the everyday vehicles, to make sure actions and choices are up to spec, or in simulators to test for more extreme events that you want to be prepared for, but never experience in reality.
HR — Hiring Better — Recruiting Staff that Match Requirements
Head Hunters keep massive collections of files of both active job seekers, and those who would consider a job, if offered, but otherwise are unprepared to move. AIs can examine candidate-records and then classify them according to personality, compatible company culture, skill sets, emotional balance, and innumerable other factors which make them a likely fit. Instead of 20 so-so fits. You could interview 1, 2, or 3 and be done in days instead of weeks of months.
Marketing — Product Pricing and Placement
Sertraline, a common antidepressant, sells for US-50¢ in Canada, but over US-$3 each for a 50 mg capsule in neighboring New York State. Where you sell determines your profits, and AIs can assess markets, developing better pricing schemes that get your product where it is needed, perhaps at a lower cost, while maintaining profit levels. In markets where it costs 6 times less, you may be able to sell six times more product because people aren’t reluctant to spend so much less to maintain their health.
Marketing — Design and Spending
People are now designing their own running shoes online, to be made, and then delivered. They fit only one person perfectly, and those people are willing to pay a premium price for that. AIs can help you to reach niche market segments, contributing both to efficient one-off manufacturing and driving interest in unique products.
Marketing — Analytics
People can say “Purple is the new black” until they are blue in the face, but they are seldom correct. Analyzing markets trends is a job for Artificial Intelligence, sweeping through fashion industry news (in this example), recognizing cyclical trends, and incorporating day to day events to find out where consumer sentiment is headed. Except for the Eureka-moment that suddenly cause major disruptions, most trands are predictable, if you have enough data.
Marketing — Target Consumers with Interest in your Products
When a customer expresses interest in your product on Social Media, that act makes them an ideal marketing target. Responding with a simple advertisement—or in a more sophisticated response, a coupon or special offer immediately—that will enhance your sales. But a human can’t do it. It takes an AI churning through terabytes of information constantly to react that fast.
PR Analytics and Response Planning
If an imaginary transportation facility refused to accept responsibility for damaging musical instruments in its care, or had three times the pet fatalities of all other companies in that line of work, it is going to be a public relations nightmare. Of course that wouldn’t really happen, but if it did, properly educated AIs could look at all of the historical instances and create a plan to rebuild public confidence.
Sales — Forecasting
Humans forecast (literally trying to predict the future), whereas AIs analyze records and precedents to calculate, when determining trends. Needless to say, AIs are much better at it since they have no biases to color their results. But
Sales — Convert Cold into Warm Calling
Cold calls for salespeople require a thick skin, cause a lot of wear and tear on their emotions, and waste a lot of time. People that are already enthusiastic about a product make everyone’s life easier. AIs can sort the wheat from the chaff in all sort of environments—particularly social media and business press—so instead of a 5% close-rate from cold-calling, results can rise to extraordinary levels by creating calling lists of interested and qualified “warm” clients.
Sales — Response Suggestions during Live calls with client
When talking to a client, we use “Active Listening” skills, but when the conversation grows long and we grow weary , we may overlook some crucial factors. An AI monitor can speak in the ear of a telephone salesperson (or even during a live conversation!) and help them ask the next “right” question to address the client’s concerns, as well as giving them the most appropriate solution for their problem.
Sales — pre-sale ChatBots
Having an AI fielding incoming calls allows them to answer preliminary inquiries, and then extrapolate additional needs by examining their online presence. This provides the perfect opportunity to follow up with an e-mail solicitation that reflects the high value of the call.
Support Deflection & Enhancement
Expert System ChatBots can address hundreds of customers simultaneously, which is good for AMAZON™, or even the local electrical utility during a power outage. Customers get an immediate answer when they call with no need to wait on line been frustrated. In the case of an electric utility, if the answer isn’t immediately available, they can still call by 2 minutes later without guilt and ask again.
Modern people don’t want to sit through a 60 second advertisement. Exposure to the high speed interactions of simply “being online” has eroded the ability to pay attention so information has to be tight, factual, and fast—if you are going to get past the five or ten-second threshold. AIs can use all the right keywords to stimulate interest so they’ll stick around for the entire message. If you don’t know what FOMO is, you need an AI to create your advertising.
Personalization of Advertising Services
The previous observation leads to advertising directed to one person, using all their personal cue-words scraped directly from their own online presence. Big or small, if you can profile your customer and use their own language, idioms, and conversational techniques, you have many times the chance to make a connection resulting in a sale.
Cashless Checkout at retail stores
See that little dot on the pile of salt? That is IBM’s newest complete computer. Sitting on the fingertip is a die composed of 64 identical computers. One dot has all the power of the 32-bit 80386-chip computer from 1985. Powered by ambient light, these could be on everything and identify what they are attached to. The AI would simply identify every item selected, all at one time, bill a customer’s account and print (or e-mail) the bill as they casually left the store, no money having changed hands.
Security of IoT Devices
We live our lives everywhere, almost constantly connected. We’re only disconnected by a deliberate act, and often not successfully. The Internet of Things (IoT) runs our lighting, opens our garage as we approach, maintains household temperature, orders our food, arranges delivery, runs the robotic vacuum when no one is home, and monitors our household security. AI makes everyday life easier.
Everything we have now has the potential to be permanently connected. Crooks have been known to hack into our household cameras and watch us in our daily lives. AIs can monitor activities and eliminate such intrusions before they succeed. As above, who wants some creep to be watching your child sleep on your baby monitor? The same AI that manages our daily lives can keep those people at bay.
Video Meeting Management — AI Directs camera to speaker
Video meetings are supposed to help us connect with others on the team. However, a trio in Seattle meeting with a group of seven in Silicon Valley and two more in Austin can’t really see each other that well. A lot of those meeting tools come with a little joystick for steering the camera so you can look at the person who was speaking. But first you have to figure out which city they are in, and then you have to identify the person speaking, all while trying to pay attention. An AI can automatically steer the camera, froe attendees can pay attention and have both hands free to write notes.
Computer & Network Analysis for Enhanced Security
We have highly effective heuristic anti-virus programs that are clever enough to recognize dangerous programs. Now imagine your whole computer system like that, intelligently monitoring and blocking compromising activities in real time. Powered by AI, we may have to rename these protective programs to anti-malice programs.
User Behavior Analysis to prevent Security Breaches
When a system is AI monitored and someone logs into a computer, they have a very distinctive pattern, unlike anybody else’s, which uniquely identifies them, and only them. It could be the speed at which they type, which keys follow in rapid succession, and which letters are always preceded by a pause. You’re the only “you”.
Financial Transaction Security through Biometrics
As above, you own body is your password. It could be the fact that if you usually type at 50 words per minute, you almost always type your own name at a rate of 150 words per minute, or more. AI analysis will eliminate the need for passwords because no one else can be you, no matter how skilled they are.
As mentioned above, people have much shorter attention spans nowadays. Customer service studies are revealed that most people get frustrated after 16 seconds of being on hold. Faced with a 45 minute delay before being able to speak Customer Service, most people actively seek ChatBots online because of the instantaneous response. Many bots are sophisticated enough to pass a Turing test (people can’t tell if they’re human or machine). It’s the way of the future.
ChatBot Analysis to enhance CRM
Customer Relationship Management (CRM) can be a significantly enhanced when ChatBot records are analyzed. This is also how AI ChatBots will be trained in the future, seeing successful and unsuccessful outcomes, learning strategies, and building upon them for future interactions.
Roboticized Service Advice (Oil Change Shop)
Almost any counter service can be managed by an AI. When a complete record of car maintenance is available, there is no wheedling or cajoling to get people to (for example) change transmission fluid unnecessarily. Your car’s own AI will let you know what you need so you can order service confidently.
Using Natural Language Interpretation will be commonplace very soon. On the phone you might not know you’re talking to an AI; in person it won’t matter because the AI will be so knowledgeable that it will be completely relatable. In Japan they are giving them human faces, too. People are really enjoying the experience and it will only get better as time goes on.
Solving problems with Deep Learning and Neural Nets helps AIs think more like people. They’re not intuitive, but can handle trillions of operations per second, allowing them to emulate our thought processes. They can find content on the net, assemble it in a way that is attractive to individual users, and deliver it. Each one is slightly different—unique—and attractive to the recipient. This can keep dynamic webpages interesting and e-mails to clients fascinating—and all this without needing to generate brand new content.
Connectivity is the watchword of this decade. Through AI Management, we have found that our older architecture, which should have been overwhelmed by the amount of data it is asked to carry, is still function fine. The AI can recognize failure scenarios, work around them, and enter maintenance requests before disaster strikes.
Provisioning Services to Maximize Throughput
More importantly, it will still be useful for another 20 years while we update it. AI automated switching sends data through less used channels, and optimizes flow during high demand periods. The estimates tell us that AI should manage this greater usage, but that peaks should only be 1 or 2% higher than current data loads.
Geophysical Image Analysis
Geophysics and the related sciences help us track new forest growth, old forest loss, advancing and receding ice sheets at the polar caps, and help us discover mineral deposits and other resources. An AI compiling all that data and combining it with satellite imaging, ocean buoy data, and world temperature records will give us the most comprehensive picture of our planet that has ever existed.
As above, AI may turn our meteorological science on its head. Who knows? Maybe we will find out that Global Warming is entirely natural and we are actually on the way to an Ice Age! It might be very revealing and exciting to see what it can teach us once it has “all the facts” in one place.
We can look at thousands of stars in a given night, but when we focus on a patch of sky which you could cover with the tip of your finger at arm’s length and find thousands of stars, the sheer size of the Universe becomes obvious. Finding planets with our new technology is easy. Analyzing all of them to see if they have biology created agents in their atmospheres—that is going to take a lot of computing power and the ability to assess huge amounts of data and recognize patterns. That’s what AI is best at!
Space Colony Design
We need a city on the Moon. Why? If the Universe sends a big rock at us, like it did with the dinosaurs, we might all be wiped out. A Lunar city would preserve our race. An AI could do the practical engineering analytics to build inside hollow lava tubes at first. Then it could tell us how to enclose a crater, pump it full of atmosphere, and then create a shirtsleeve environment. Until we live on two different bodies, our race is at risk.
Online gaming is great, but sometimes there are a limited number of competitors. AIs can take up the slack, filling in for players so the game can continue. The AIs use human-derived moves, and so look entirely indistinguishable from living players.
Artists want to create art. Soon we will all want to have customized objects in our homes. If we don’t own a 3D printer, we’ll send an image to a 3D printer company and they will create the result for us. If the object is incredibly ordinary, requiring little “art” people might not want to tackle it. An AI can run a 3D printer to create all the banal items like mixing bowls for the kitchen, a replacement cap for your windshield washer fluid, or a flyswatter. Somebody has got to do it!
AI is absolutely essential for this process. What else needs to be said? Ask your Tesla Dealer and you’ll discover that all the automation hardware required for a Self-Driving vehicle has been installed 2016!
Self-Driving Transit Systems
Once we replace cars, next on the list will be buses, subways, streetcars and so on. AI run systems will be more efficient, less costly, and on-time, all the time.
Commercial Delivery of Goods
Say goodbye to tractor-trailer rollovers on exit ramps, or overworked drivers falling asleep and running off the road. And we’ll all breathe a sigh of relief when dump truck drivers are relieved of the responsibility of putting the box back down after they’ve delivered their load. No more highway signs destroyed; no more bridges damaged; no more power lines ripped down plunging thousands into darkness. AI can’t come too soon!
Once our roads are safe, our rail systems can be automated. No more trains stopped across a level crossing for 40 minutes because of bad scheduling. No more over-speed entries into curves, or around blind corners, only to discover cars stopped across the tracks. Bad sections of track can be reported for maintenance before something derails. Again, AI can’t come too soon!
The truth is that most commercial aircraft are almost fully automated now. It’s a small step to reach the point where the pilot never touches a control from takeoff to landing.
Automated Sea-going tankers
Goodbye and good riddance Exxon Valdese. These behemoths, like aircraft, are so fully automated that they can pilot themselves. Even the ports they visit supply their own pilots to get into harbor and docked.
Autonomous Manufacture-to-Order from Dresses to Vases
Unlike 3D printing, companies that make clothes can use AIs to custom cut material to your exact size, in precisely the material you desire. It can be sewn, boxed, shipped, and delivered without a human hand ever touching your item before you do.
An AI can track homelessness and poverty, and come up with solutions that may seem odd or unexpected. Unused public buildings such a retired recreation centers, closed schools, or properties owned by a municipality for tax arrears are often suggested, but interestingly, there are five vacant homes in the United States for every homeless person. AI will solve this problem in a way we will find surprising.
Basic speech recognition has been with us for some time. This entire article was created using speech recognition and so far has offered no difficulty even with scientific terms. It uses contextual relationship filtering to figure out which homonym to use (they’re, their, there/where, ware, were, we’re, and so on). It’s not AI yet, but soon it will be utterly conversational, eliminating Siri’s canned responses in favor of legitimate insights and truly useful advice.
NLP Language Interpreters
The Star Trek fans among us will be excited about the Universal Translator (at least for Earthly languages). Google already offers remarkably intelligent translation in over 100 languages. With a fully integrated AI the translation will be so reliable that misunderstanding will be much more difficult to achieve.
Visual Guide for the Blind
Blind people will have an option to have the world around them described. Whatever is nearby, in front of them, or posing a threat from any direction, the AI can keep them apprised of their environment increasing safety and allowing them to interact more freely with the world.
Aural Guide for the Deaf
Called the “Invisible Handicap”, not being able to hear is problematic in certain situations. An AI could monitor the local environment and transmit the information perhaps by a bone conduction device, a visual aid (smart watch or phone) or simply a mechanical alert to danger.
Sign Language Interpretation for Machines
Gesture-recognition utilizing AI could translate sign-language to verbal, caption, or video depending on need or circumstance. There are dialects of sign language, depending on region, as well as slang, but this should present no serious challenge to an AI trained in most of the different styles.
Automated Service Provider
Service clerks that are not required to leave their station in order to perform their duties, such as a desk attendant and a Self-Storage Unit, could see their jobs turned over to Artificial Intelligence. That it should come as no surprise because any job which is filled with repetitive activities that can be handled by machine will probably end up being handled by a machine. Of course there’s an opportunity for maintenance and repair duties, but just sitting at the desk will probably vanish.
Quick & Dirty Apps with functionality defined by pre-built modules should be remarkably easy for an AI to design. As it becomes more adept and sophisticated, more complex things will become possible.
The above means that this will be necessary, too. While developers work on their software, and AI knowing the desired outcome can build an ergonomic User Interface so that the buttons and action-objects are in the correct place for ease of use.
Naturally, the User eXperience is an important consideration, too. Youths might like purples and Neon colors for personalization, but the functionality of the application, the ease of navigation, the fast efficient changes without clunky operation, is going to be the big selling feature and is vitally important. AIs can look at successful and failed designs in order to learn to make optimal choices.
A program that evolves is more useful than one that has a fixed or limited utility and can be outdated. One which can learn, that can find new ways to solve old problems, will remain useful for much longer. We can do away with the planned obsolescence concept of producing a new phone every year in order to be up to date.
The idea of 30 students sitting in a class, facing a teacher who is anxious to meet all the requirements for the semester, can slowly fade away. The idea of a number of learners, approximately the same age, meeting in order to gain an education is socially useful. With an AI in charge however, it becomes less necessary for everyone to advance at the same pace. The AI can tailor the lessons to each individual, depending entirely on their own progress. Falling behind or getting ahead would be students proceed at their own pace under the guidance of an artificial intelligence.
Education Teaching Online
Online and MOOC (Massive Open Online Courses) is a great way to do an end-run around the high cost of education since courses are generally provided for free. Most progressive countries fund or at least subsidize their youth for additional education. Smarter graduates, equals a better workforce, equals a better country. Like in-person teaching, mentioned above, students proceed at their own pace under the guidance of an artificial intelligence.
If someone is unable to grasp a concept, who better to explain it than an AI with so much experience in analogical analysis? Analogies are some of the best tools for getting a concept across. AIs may be digital entities but they are experts in analogy.
Online Political Voting with Biometrics
One person—one vote, or so the saying goes, but sometimes the turnout is lower than anticipated. People don’t feel like they can make much of a difference, but what if everyone was uniquely identifiable? What if you could vote online, from home, or from your phone because an AI could be certain of your identity? What if your vote was subtracted from the vote total cast by your representative? If you agreed with the stance, let them vote. If you objected cast your own vote. If enough people don’t like the stance, you could defeat your own representative who wasn’t representing your view.
An AI can interpret how much you make, how much you spend, what your savings goals are, how tolerant you are of investment risk, and then run your finances for you, paying your bills, putting aside an emergency fund. Holding back some cash for that all-inclusive trip to the Cuban resort… You could just cruise your way into retirement, probably a lot earlier than most others who don’t use an AI program…
Google is undertaking the digitization of the existing 130,000,000 unique books on our planet, using Optical Character Recognition (OCR) and little vacuum-powered page turners. Once it is complete everyone can look up anything that was ever written. Law students won’t have to go through dozens of volumes to find precedents; they can just place an enquiry and find what they need in seconds. Imagine the industry of law becoming fast, efficient, and accurate.
Network Honey Pots to trap malevolent Hackers
AIs can make a honeypot (bait for hackers) look like an intrinsic vital, functioning part of the system. It doesn’t just sit there; it processes data constantly (albeit useless data) to make sure it looks absolutely essential. Back-tracking a hacker while they think they are succeeding in their hack can result in arrests. Hurry AI, we need you now!
Finally, Facebook has undertaken the task of profiling its members who show characteristics of potential suicide attempts. They are working to get them help, a laudable goal.
Where does it end? Well, honestly, it doesn’t. Everything you can imagine can have an AI component, and most will have it within the decade, though there I no shortage of development ideas right now and early implementations. Soon we’ll need AIs to help us design AIs… such as…
AI can help in dynamic pricing for retailers. As retailers fight their competition to provide better pricing to their customers, AI can help optimize the pricing keeping in view different factors like profit, sales volume, inventory, etc.
Expert System Design
AIs will be most useful as consultants for humans. The Scientific Renaissance is most thoroughly over. No one person can have a broad enough knowledge covering all areas of science. There is now simply too much information. A single Expert System can absorb all the knowledge of 1,000 farmers and be able to grow perfect crops, but be completely stymied by the how to treat a sick whale, or conduct an Icelandic wedding ceremony. It will take an AI System to integrate all the smaller AI Expert Systems, and then the limits to what we can learn are unfathomable!
We have remote robot surgeons now that can handle incredible detail work.. An expert system based on surgery and physicians’ real experiences could provide faultless surgery, able to cope with any incident that any of the contributing surgeons had encountered.
We might desire to learn about Water Furnaces and the AI could provide an AR representation large enough to walk around inside so we could truly understand how it worked. This same principle would apply to brainstorming session in industry for maintenance, In product design, or any other consultative process.
Shooting Rampage Prediction (via Social Media)
Autonomous Network Threat Recognition & Response
Antedated Beauty Contests could use AI instead of Judges
Micro-Drone Pollinators to help crops
…and more than you could possibly imagine…
It is hard to decide where to stop. There are hundreds more of these utterly practical possibilities. If nothing else, it has given you a small glimpse into the possibilities.
Just remember, although intrinsically innocuous, if we fail to implement AI properly it will come back to haunt us. We need to use foresight to make sure that these and any other implementations are an extension of humanity and not a self-interpreted replacement for it.
A hundred years from now, our descendants will look back on this age and say “We are so glad that you got it right!” because, if we don’t, the alternative is that there will be no one to look back.
Artificial Intelligence is not “Skynet” from the Terminator movies. If we proceed thoughtfully and logically we will enter a New Age for humanity, and our progress will race forward in leaps and bounds.
If we do it wrong we may have to take the old science fiction joke about why there aren’t millions of other civilizations in the galaxy seriously. The punchline to that joke was “It looks fine to me. Go ahead…push the button”.