Artificial Intelligence (AI) is waking up and knocking on your door. Almost 60 years in the making, the AI winter — a time with less research and funding allocated toward AI advancements — is over. With 83% of surveyed businesses noting AI as a strategic priority and 75% stating AI as the key to finding new business and ventures, Artificial Intelligence is no longer a product of science fiction. From detecting fraud to efficient hiring methods and predicting machine maintenance, Artificial intelligence is quickly becoming an innovation driver in both the workplace and everyday applications.
Ready to learn more? In this article, we’ll outline the evolution of AI, detail how it works, and help you make a case for AI in your organization.
Artificial intelligence: 60 years in the making
Artificial Intelligence is a branch of computer science which enables computer systems to learn and perform tasks normally associated with human intellect, such as speech recognition, decision-making, and visual perception.
In the 1950s, the reality of AI was at best, out of reach — computers couldn’t store or execute information and costs were astronomical. Then, mathematician Alan Turing asked a simple, yet life-altering question, “can machines think?”. The answer, a rousing yes, changed the course of history.
Between the 1950s and 1970s, the computer industry found its footing with computers performing faster, becoming more accessible, and less expensive. A 1970 Life Magazine article reported that in a mere three to five years, machines would soon have the same intelligence of a human being. However for that to happen, major advancements in storage capacity and computational power were needed.
In the 1980s, two important techniques were developed. The first, “deep learning,” allowed computers to learn by way of experience. The second, the “expert system” mimicked human’s capacity to make decisions. Computers began to use reasoning based on “rules” — primarily an “if-then” process used to answer questions.
Fast forward to 1997, Dragon Systems developed and implemented speech recognition software on Windows. The 2000s brought speed and storage options, like the cloud, catapulting the use of computers into the mainstream and giving AI its turn in the spotlight. Today, AI is on a fast track due to three major industry improvements.
- Graphics processing units (GPU): Demand within the video and gaming world resulted in the development of improved and less expensive GPUs — a component needed to enable and build AI software.
- Big Data: The algorithms used by AI are essentially “trained” by the vast amount of information in big data. Those algorithms then help AI to process that information at an incredible pace and make data accessible and more usable.
- Algorithms: Algorithms allow the automation of tasks that were once thought only possible with human intelligence. Algorithms continually improve and become more complex through the use of layers with hidden variables that sort through and optimize results.
Today, artificial intelligence, or machine intelligence, AI-enabled machines can perform the following acts:
- Speech recognition
- Problem solving
- Ability to manipulate and move objects
Fundamentals of Machine Learning now.
Leveling up artificial intelligence: ANI, AGI, and ASI
There are three types of AI: artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial super intelligence (ASI).
- Artificial narrow intelligence: ANI is categorized as weak AI because it only specializes in a narrow range of parameters or situations, like speech recognition or driverless cars, for example.
- Artificial general intelligence: AGI is considered a strong AI as it works on a higher-level that’s in line with human intelligence.
- Artificial super intelligence: Although this type of AI is not currently developed, ASI means a machine has superintelligence or is smarter than a human.
ANI is the most simple type of AI to identify and is already being used by most individuals. Here are a few examples of ANI you may recognize:
- Self-driving cars: Driverless cars have no steering wheel or pedals, run on 3D maps, and are controlled by a Google Chauffeur.
- Voice-activated devices: From Siri to Alexa, most consumers own and use ANI on a daily basis through smart devices.
- Mail filters: All, if not most, email inboxes give users the capability to sort spam and flag special messages.
AGI takes AI one step further, requiring a machine to perform human-level intellectual tasks. However, in order for a robot to be classified as AGI, it has to pass a few tests that check the machine’s intelligence.
In particular, the Turing Test checks the ability of a machine to act like that of a human. If a machine scores a 70% or higher, it is classified as an AGI robot. Another analysis of AGI compatibility is the Coffee Test, requiring a robot to enter a home environment, find coffee, and learn how to brew it. An AGI prospect must also pass the College Robot Test, which engages a robot to enroll in school and successfully take classes. Finally, a robot can take an Employment Test, where it is must pass vocational tests including driving and writing exams.
How does AI work?
Now that we understand how AI has evolved over the past 60 years and have an idea of the fundamentals, let’s look closer at how AI works.
Two fundamental techniques allow AI to work effectively, Machine Learning (ML), and Deep Learning (DL). ML and DL create the environment for AI to function. ML is a computer’s built-in tools and algorithms that “learn,” or use existing data to produce accurate predictions. DL is an automated feature within Machine Learning that allows a machine to notice patterns and sort information into categories, allowing it to “think.”
Let’s look more closely at some of the technologies within AI and how they can work for your enterprise.
- A machine learning platform can use information from a number of different data sources — like development and training tools, along with other algorithms — to predict and classify information.
- Deep learning is a machine learning technique that uses pattern recognition and classification in order to work with large data sets.
- Neural networks are a technique of machine learning using statistical algorithms that are fashioned after the behavior of neurons in the human brain.
- Cognitive computing is a type of computing that uses reasoning and high-level understanding. It is not considered machine learning because it uses a combination of several AI technologies to produce results.
- Computer vision enables computers to act and process like the human eye. It analyzes the context of digital images and videos to produce numeric or symbolic information for decision making.
- Natural language generation simply means generating text from computer data. This process is often used for customer service, reports, and business intelligence summaries.
- Graphical processing Units (GPU) are part of an electronic circuit that accelerates the creation of images on a display device. GPUs are imperative for AI to work successfully.
- IoT, or the Internet of Things is a network of connected devices that generate and share data, like appliances, smart speakers, wearable technology and medical devices. AI depends on data from these devices to generate important business intelligence.
- Advanced Algorithms are complex algorithms that are constantly being developed and combined to provide ongoing intelligent processing.
- APIs, or application programming interface, is technology that is sometimes used to access AI services. Likewise, AI uses API data flows to help businesses make sense of data that can not always be seen.
O’Reilly Report: The Internet of Things Market now.
Real-life AI applications
Chances are there is more AI in your life than you realize. Have you interfaced with a cyber assistant on your favorite shopping site? Asked Facebook for assistance creating an ad for your business? Summoned Alexa to play your favorite playlist? Said hey google to find your way home? These are only a few of the ways businesses are using AI to make life easier and provide more to customers. Here are a few more:
- Amazon offers transactional AI with algorithms that continuously become more advanced. Currently, it can predict your purchasing habits and provide product information before you even know you’re in the market for a specific product.
- Pandora’s musical DNA process uses more than 400 musical characteristics from songs that have been manually analyzed by professional musicians to recommend new songs to users based on their preferences.
- Nest is a thermostat that can be voice-controlled by Alexa to learn and activate your heating and cooling preferences.
What about cloud-based AI? With its scalability for growth and resource access in real-time, AI with a cloud infrastructure is emerging as a leading option. Platforms currently available include Microsoft Azure Machine Learning, Google Cloud Prediction API, TensorFlow, Rainbird, and Meya, to name a few.
AI business benefits
AI adds value to existing computer capabilities by continually and reliably providing computerized tasks and information. AI can help improve your bottom line through:
- Automated processes and tasks
- Fewer mistakes and human error
- Increased productivity and operational efficiency
- Improved business decisions with access to real-time data
- Improved data learning with increased access to large data pools
- Improved service with customer knowledge
- Quality lead generation
Moving into specifics, here are five detailed ways AI can work for your organization:
- Data collection and analysis: AI makes data collection and analysis affordable, intuitive, and timely so you can automatically learn more about your customers and secure repeat and new business.
- Smarter hiring: Machine learning algorithms can determine best practices for your specific hiring needs and create a short list of the best candidates.
- Back office efficiency: AI can handle tasks like accounting, scheduling, and other data-to-day functions in a snap, without error.
- Customer service: Virtual customer service assistants are working 24/7 and can provide help to your existing and potential customers without the supervision of a human.
- Targeted marketing: Sorting through and categorizing all available data about your product or service is an AI specialty — allowing you to focus on marketing that specifically pinpoints the needs of your customers.
Guide to Machine Learning now.
An AI-integrated ecosystem
More and more businesses have AI workloads overseeing important business processes. Finding a platform solution that works specifically for your business and connects a range of third-party applications is an important part of the IT infrastructure. Talend Cloud can seamlessly prepare and integrate your cloud and on-prem applications and data, reinforcing your AI ecosystem.