Introduction to Artificial Intelligence
Over the past few years, artificial intelligence has reached a tipping point that spans across industries. Colleges and universities have begun offering courses and degrees in AI to train students for jobs in this emerging field. Industry research shows that most leading software vendors have committed to embedding AI and ML technologies into their respective platforms and solution sets. Additionally, in February 2019 President Trump signed an Executive Order promoting the development and regulation of artificial intelligence in the United States Even though AI is considered an Early Adopter technology on the Technology Adoption Curve, it is garnering a lot of attention, and it holds significant promise to offer personalized and adaptive services to constituents that can ultimately improve student outcomes.
Figure 1: Technology Adoption Curve
Source: http://www.caroli.org/the-technology-adoption-curve/
The concept of AI originated in the 1940’s with the concept of an artificial brain and artificially intelligent robots. Prior to 1949 computers could execute commands but they could not store commands. By the early 1950’s computers advanced to the point of being able to store commands and artificial intelligence research was founded as an academic discipline in 1956. In the 1960’s scientists began developing machine learning in robots. For several decades true success with AI technology was deterred by the fact that computers lacked the computational power to store large volumes of data and process it fast enough. With the exponential increases in computer processing over the past two decades, AI has resurrected and is now being imagined into the realms of AI-based voice activated expert systems, robotics, data mining, search engines, etc.
Artificial intelligence is an emerging ecosystem of technologies that has become an umbrella term that represents a myriad of technologies, such as machine learning, [CMB1] deep learning, computer vision, natural language processing (NLP), and machine reasoning. The evolution and addition of new AI solution types indicates how vast and sophisticated the realm of artificial intelligence really is. While the terms AI, DL, and ML are often used interchangeably, they each have unique roles and capabilities. As shown in Figure 2, machine learning and deep learning are often referred to as subsets or cousins of artificial intelligence. The key contribution of AI from a systems perspective is that AI is merely observing all the behaviors of the environments in which it is placed. These observations are then put into a computer program and passed off to machine learning. When enough observations and machine learning iterations occur, the methodology of deep learning takes over.
Figure 2: Technology Adoption Curve
Source: towardsdatascience.com
The combination of AI, ML, and DL is allowing companies to combine aspects and deliver applications such as Chatbots and AI assistants. These applications are easily embedded as a code set within a website. When a person logs into the website, the Chatbot AI component can observe what the user is doing on the website. After instant observation, the machine learning component can then trigger action to begin interacting with the user. Behind the scenes, the combined machine learning and deep learning begin analyzing and predicting how the experience with the user can be intensified, such as recommending certain purchases based on previous purchases.
Uses of Artificial Intelligence
Even without realizing it, we experience artificial intelligence and machine learning in our day-to-day lives including:
· Voice recognition tools such as Siri, Amazon Echo, and Google Home
· Recommendations presented to us on Netflix or Pandora
· Email filtering in Gmail which sorts our mail into “junk,” “primary,” or “promotions”
· The predictive text you may see on your mobile device when responding to a text message
· Predictive text that pops up during a Google search
· Product recommendations on Amazon
· Fraud detection messages sent to you by your financial institution if a transaction appears to be out of the ordinary
· Uber utilizes AI to determine how long it will take you to get from your location to your destination.
Most industries have incorporated artificial intelligence for a variety of reasons including, but not limited to automating work, improving efficiencies, saving money, and performing undesirable job duties.
Healthcare
Surgeons are beginning to utilize robots to assist them with surgeries to improve accuracy. Electronic Health Records have also incorporated artificial intelligence to sort through and find historical patient information faster.
Education
Educational institutions are using artificial intelligence to improve student success and engagement. A few examples include chatbots for students to ask questions any time they need, including after hours and AI assistants to help students while allowing teachers to focus on more meaningful work.
Retail and e-commerce
Artificial intelligence can provide automation using the large amounts of data, helping decision makers with decisions related to customer data, stock, and inventory. Additionally, chatbots provide customer service by answering customer questions.
Banking
Artificial intelligence can assist with security by analyzing spending behaviors and spotting abnormal behavior and fraud.
Energy
AI is being used in the energy sector to provide accurate energy forecasts, improve energy efficiencies, and to help consumers save money.
AI Platforms
There are several companies offering AI tools to help improve operations. In addition to pre-built AI tools, there are AI platforms that allow individuals and organizations to build their own AI solutions. Using an AI platform requires technical resources, but not necessarily a full team of developers. Choosing to use an AI platform will allow you to customize and build your own solutions and give you a little more freedom than implementing a pre-built AI tool. Utilizing an AI platform may require some staff to receive specific training with a particular platform.
It is essential to consider the whole enterprise (all systems used across an organization) when selecting AI solutions. Implementing pre-built AI tools may be less work, but they may only benefit limited areas within a company. When choosing pre-built solutions, you also need to consider if and how they can integrate with systems currently being used.
If you believe that AI will be a significant component in our organization in 3 to 5 years, then you should take into consideration how you can integrate AI, ML, and/or DL to benefit the entire company in order to be connected and operate seamlessly. Deciding to utilize an AI platform will also affect your staffing decisions. It will be important to train your current staff and hire the right talent.
Concerns With AI
One of the biggest concerns with artificial intelligence is security and privacy. The need for people to have a sense of privacy on campus will impact how far many institutions will be able to go with collecting data through AI, ML, and DL. Therefore, it is critical to understand how AI, ML, and DL work together to produce the AI experience.
Although the technology has advanced drastically over the last several years, there are still some concern whether it is where it would need to be to take the place of humans. For example, some chatbot experiences have been very positive, while others have not been as positive due to the responses being too vague or even wrong. It is important to consider the scope of work that you expect the AI-enabled tool to perform, and fully understand the product options and capabilities.
Conclusion
It is fair to state that AI has made it past the hype-cycle. However, what has made AI valuable is the evolution of machine learning and deep learning as subsets of the overall impact and effectiveness of AI.
With so many institutions trying to understand the personal learning styles and student success factors, AI is the perfect solution with ML and DL as subsets. As great as this may all sound, it will be critical to monitor, manage and facilitate where and when solutions with built-in AI may be ready for prime time.
With numerous technology vendors now incorporating AI features in their product as a competitive advantage, it will be imperative for CIOs to be aware of which vendors are offering a free vs. paid solution. It will also be critical to collaborate with others to validate that the AI solutions will seamlessly integrate into their current environments.
Comments