Why Learn AI

To have a meaningful discussion regarding Artificial Intelligence (AI), one must first understand what exactly Artificial Intelligence is. What does it aim to solve? What are its constituents? When was AI invented? What kind of research and development has gone into it? What are the current and projected trends surrounding the technology? Let's try answering all of these questions.

AI stands for Artificial Intelligence. According to Merriam-Webster, the dictionary defines Artificial as "humanly contrived often on a natural model"; and Intelligence as "the ability to apply knowledge to manipulate one's environment or to think abstractly as measured by objective criteria (such as tests)." That has always been the aim of Artificial Intelligence. To gain meaningful insight into why one should learn about Artificial Intelligence and its close cousin Machine Learning, one must understand its history.

Machine Learning is essentially a subset of Artificial Intelligence.

Brief History and Inception

AI has been around for decades. Research on AI dates back to the early 1950s, with the research centered on symbolic functions. The US Department of Defense (DoD) began training machines to mimic human logic and reasoning in 1960. The Defense Advanced Research Projects Agency (DARPA) started a program for integrating AI technologies into a cognitive assistant in 2003, developing an AI project called Cognitive Assistant that Learns and Organizes (CALO).

Market Trends and Analysis

According to Gartner, during 2018-2019, deployment of AI in organizations grew from 4% to a whopping 14%. The following infographic identifies the AI technologies that will be on a CIO's radar, contributing towards a massive business impact in the next five years.¹

According to Forbes, AI and machine learning are poised to create an additional $2.6 trillion in value by 2020. Major business impact is expected to occur in marketing and sales, up to $2 trillion in manufacturing and supply chain planning, and a worldwide expenditure of $77.6 Billion in 2022 on cognitive and AI systems.²

The demand for jobs is shifting away from unchanging tasks and towards those that are cognitively driven and require digital skills. According to an analysis of historical trends of new jobs created to old jobs and an adjustment for a lower labor-output ratio due to the labor-saving nature of AI via smart automation in a report by McKinsey ⁴ , new jobs driven by AI investment are poised to augment employment by 5% by 2030.

Current use cases

AI technology covers a huge breadth of use cases across industries. Most AI implementations consume interaction data combined with historical data in real-time and draw analyses from them.

Retail

The retail industry has started applying AI, machine learning, and automation in the supply value chain. The three areas with the highest impact are promotions, assortment, and replenishment. E-commerce companies are leading the way using AI to forecast trends, optimize warehousing and logistics, fix prices, and personalized promotions. The impact of such an AI-enabled prediction is already visible in the market. Otto, a German e-commerce merchant, has cut surplus stock by 20% and reduced returns on products by over 2 million returns per annum. This has been achieved by exploiting deep learning to analyze billions of transactions and predict what customers may buy before an order is placed.

Autonomous robots play a major role in warehouses, especially in cost reduction and eliminating injuries. Various manufacturers, warehouse owners, and outlet stores have reported a reduction in stocking time after using autonomous guided vehicles, trolleys, and electronic beacons. These include Swisslog (30% reduction) and DHL. Carrefour and Target have also been able to use the beacons to collect their customers' movement data to map customer behaviors and purchasing patterns. These trends have been fed into AI algorithms to determine which personalized promotions to send customers as they shop.⁵

Strides in computer vision have allowed AI to identify desired goods by taking a picture or via preference patterns from images and videos liked online by consumers. Smart speakers and other home assistant devices utilize AI algorithms to track customers' purchasing histories and set reminders or even order utilities for them. For example, Google Home is partnered with other brick and mortar retailers like Costco, Whole Foods, and PetSmart. Amazon's Echo Look incorporates a camera into Alexa's virtual assistant function and recommends styles based on the user's wardrobe and body shape, combining AI and computer vision.

Healthcare

The healthcare industry generates and processes a massive volume of information surrounding patient histories, medical images, epidemiological statistics, and other data. As such, AI can draw inferences and recognize patterns from such massive volumes. AI combined with digitization and visualization can help predict the spread of diseases, assist doctors in diagnoses, transform chronic disease treatment, and allow patients to be monitored remotely.

Johnson & Johnson and SAP have used AI and machine learning to forecast inventory levels, product mix, prices, and customer demand. Other use cases include using AI and machine learning algorithms to predict the probability of a patient being re-admitted to a hospital (CareSkore). Turbine uses AI to design custom cancer treatment regimens by simulating millions of experiments, modeling cell biology at a molecular level, identifying complex biomarkers, and the best drug to use for tumors.

Manufacturing

AI is set to disrupt manufacturing on a large scale. This includes everything from virtual assistants to advanced robotics. More specifically, AI will enable the industry to optimize processes in real-time by leveraging rapid growth in the volume of data. This would eventually lead to shortened development cycles, improved engineering efficiency, fault reduction and prevention, improved safety via automation of risky tasks, reduction in inventory costs through better supply and demand planning. Finally, AI-assisted processing can provide an increase in revenue with better sales lead identification and price optimization.

AI helped an aerospace manufacturer reap €350 million in savings. Analyzing data from the assembly process and refreshing the standard operating procedures lead to 60% of the savings. The rest was made possible by using autonomous vehicles and robots. A semiconductor maker reduced its material-delivery time by 30% and increased its production yield by 3-5% by using machine learning to propose the best time to leave the office or warehouse.

Electric Utility

The electric utility sector currently suffers from an aging critical capital asset, uncertain supply and demand, price deregulation, cost pressures, non-linear power loads, and a complex network of stakeholders and assets. There is an enormous potential for the deployment of AI technologies in the electric utility sector. AI can help forecast supply and demand, real-time balancing of the grid, downtime reduction, yield maximization, and improved end-user experience. Grid modernization is already underway with the deployment of smart meters to forecast and optimize load dispatch.

The UK's National Grid, along with DeepMind, is seeking to predict supply and demand peaks by using weather-related variables and smart meters as inputs. This could reduce national consumption by 10% and increase the use of renewable power. ⁹

"Digital wind farms" are a concept by GE Renewables using AI, which consumes turbine sensors data to optimize yields and customize turbines to varying conditions at each installation site. These conditions include past performance of turbines, real-time communication with other wind farms, and changes in wind velocity. This is expected to boost energy generation by 20%, creating $100 million in extra value over the lifetime of a 100-megawatt farm.⁸

AI can also assist in tackling non-technical energy losses, such as electricity theft. In Hungary, Eon AI was used to narrow down the list of suspicious users and helped reduce the theft by 30%. This was possible by analyzing customer data like usage patterns and payment history and comparing it to known irregular behavior. AI applied to meter data can help extract energy profiles of the largest appliances and visualize per device contribution to the electricity bill. This can help consumers by providing detailed real-time insights on their energy consumption.⁵

Education

Artificial Intelligence will play a key role in better connecting education systems and labor markets. A prime example of this was seen in 2015 when LinkedIn acquired the educational website Lynda.com. LinkedIn leveraged AI to offer a personalized online class selection for members considering a new job or career.¹⁰

Subsets of AI, namely, Natural Language Processing (NLP), character recognition, and deep learning (DL), could assist teachers in answering common student queries. One such experiment was conducted in 2014 when a Georgia Tech professor, along with his team, programmed a teaching assistant. This artificial assistant responded to students' online questions for five months without the students noticing.⁶ It is estimated that by 2017, the assistant will be able to answer 40% of the students' questions. In 2012, the University of Akron, in partnership with Kaggle, hosted a multi-vendor evaluation. This included the assessment of approximately 16,000 essays by both grading software and teachers. 85% of the time, computers matched the grades of human teachers.⁷

Security

AI and machine learning can identify and neutralize previously unseen cyber-threats, transforming cyber defense. According to a report by Capgemini, 69% of enterprise executives surveyed felt AI would be essential for responding to cyber-threats. Telecom led all other industries, with 80% of executives counting on AI to fill the gaps in defenses.

BluVector was recently acquired by Comcast in hopes of leveraging Artificial Intelligence and machine learning to detect and analyze increasingly sophisticated cyber-attacks. This was eventually packaged and launched as a service, developed in partnership with Cujo AI, called Xfinity xFi Advanced Security, which automatically provides security for all the devices in a customer's home that are connected to its network. Cujo AI is an El-Segundo, California based start-up that developed a platform to spot unusual patterns on home networks and send Comcast customers instant alerts.

Shift (based in Paris), has developed algorithms that focus narrowly on weeding out fraud in insurance by spotting patterns in data like contracts, reports, photos, and even videos. Shift with over 70 clients has amassed a huge amount of data that has allowed it to rapidly fine-tune its AI resulting in more efficiency for insurance companies and a better experience for customers, whose claims are processed faster.

Conclusion

AI is poised to cause major disruption across most industries. This is echoed by executives, 62% of them believe they will need to retrain or replace more than 25% of their workforce between now and 2023. This sentiment is more common in the US (64%) and Europe (70%) as compared to the rest of the world (55%). It is felt more acutely in larger companies, especially 70% of the executives at companies with more than $500 million in revenues see technological disruption over the next five years.⁵

Convinced and ready to kick start your journey in learning the fundamentals of Artificial Intelligence and Machine learning? To stay abreast of the wide-sweeping changes cutting across industries, it is a good time to invest in upskilling yourself, especially in a field that is poised to boom further. The Introduction to Artificial Intelligence and Machine Learning course offered by Cybrary is a good place to get started.

---

References

1. "Top Trends on the Gartner Hype Cycle for Artificial Intelligence, 2019." Gartner, www.gartner.com/smarterwithgartner/top-trends-on-the-gartner-hype-cycle-for-artificial-intelligence-2019/
2. Columbus, Louis. "Roundup Of Machine Learning Forecasts And Market Estimates For 2019." Forbes, Forbes Magazine, www.forbes.com/sites/louiscolumbus/2019/03/27/roundup-of-machine-learning-forecasts-and-market-estimates-2019/#482bd8fc7695
3. https://www.capgemini.com/wp-content/uploads/2019/07/AI-in-Cybersecurity_Report_20190711_V06.pdf
4. Pablo, Illanes, "Retraining and Reskilling Workers in the Age of Automation." McKinsey & Company, www.mckinsey.com/featured-insights/future-of-work/retraining-and-reskilling-workers-in-the-age-of-automation.
5. https://www.datasqlvisionary.com/wp-content/uploads/2018/06/Artificial-Intelligence-The-Next-Digital-Frontier.pdf&sa=D&ust=1585338041692000&usg=AFQjCNF6soRwc5-Vc3uJx3703jEnSGFCzQ
6. "Georgia Institute of Technology." Artificial Intelligence Course Creates AI Teaching Assistant, www.news.gatech.edu/2016/05/09/artificial-intelligence-course-creates-ai-teaching-assistant.
7. Simon, Stephanie. "Robo-Readers: the New Teachers' Helper in the U.S." Reuters, Thomson Reuters, www.reuters.com/article/us-usa-schools-grading-idUSBRE82S0ZN20120329.
8. Tomas Kellner, "Wind in the Cloud? How the Digital Wind Farm Will Make Wind Power 20 Percent More Efficient," https://www.ge.com/reports/post/119300678660/wind-in-the-cloud-how-the-digital-wind-farm-will-2/
9. Thomas, Nathalie. "DeepMind and National Grid in AI Talks to Balance Energy Supply." Financial Times, Financial Times, www.ft.com/content/27c8aea0-06a9-11e7-97d1-5e720a26771b.
10. "Press Releases." Lynda.com, www.lynda.com/press/pressrelease?id=4563.