September 2019:

AI Is Changing Industries And Society

Robotics and artificial intelligence have permeated life and business, creating economic opportunities as well as socioeconomic challenges.

As robots become cheaper and more capable, they’re moving from their traditional roles on factory floors into new industries. Collaborative robots or “cobots” too are being integrated into more types of environments and performing more varied tasks—from stocking shelves and assisting factory workers to performing services in the home. In industries like retail, manufacturing, agriculture, healthcare, and beyond, robots and cobots, automation, and AI (artificial intelligence) technologies are having a tremendous impact on the enterprise, often creating opportunities for organizations to achieve cost savings and helping industries overcome human labor shortages.

However, these emerging technologies are also creating socioeconomic challenges like nothing society has seen up to this point. So, how is AI changing industries, and how will these changes impact workers? Perhaps even more poignantly, is society ready to ask the tough questions, like what does it mean to be human?

AI Drives Economic Opportunities

Artificial intelligence has permeated the business world, along with the rest of society, and it is a significant driver of economic opportunity. “(AI) has introduced new occupations—e.g., Uber driver—and (a) genre of well-paid specialists who develop AI and provide security for online systems,” says Pauline Davies, professor of practice at the Hugh Downs School of Human Communication at Arizona State University. “All large companies are investing in AI. AI makes banking decisions, supply chain decisions, and productivity decisions, employment decisions, criminal justice decisions. Reliance on AI can be found everywhere.”

The economic opportunity is estimated to reach the trillions in the coming years thanks to its value in improving customer experience, generating new revenue, and reducing costs. Gartner predicts the business value created by AI will reach $3.9 trillion in 2022, up from $1.2 trillion last year.

McKinsey Global Institute similarly predicts AI and machine learning will create trillions in value for market segments like marketing and sales, manufacturing, and supply chain planning. McKinsey’s research further suggests the automation of labor could add up to 11% (around $9 trillion) to the global GDP (gross domestic product) by 2030, and the innovation in products and services prompted by AI could add 7% (around $6 trillion) by the same year.

Industries like retail, transportation, and manufacturing have changed and will continue to change as a result of AI and IoT (Internet of Things) technologies. “It’s hard to think of an industry that won’t be affected,” Davies says.

Global AI-Derived Business

Decision Automation Growth
2018-2022

Source: Gartner

…It’s hard to think of an industry that won’t be affected…

“Our shopping experience has already changed dramatically with Amazon and online retailers and our choice of products is being influenced by adverts selected specially for us based on monitoring of our online browsing and shopping activities, the news stories we read, and other routine data-gathering activities by the big tech sector. Transportation is changing right now with autonomous cars. Soon, goods will be ordered, manufactured, and delivered with minimal human intervention. Factory, warehouse, and driving jobs will disappear, as will jobs in agricultural production. The role of professionals such as lawyers, doctors, and accountants will change as AI analyzes documents, diagnoses sickness, and scours tax law. Even fashion models will be affected, since AI-generated images in online advertising fulfill their roles.”

In manufacturing, Davies says robots and cobots can produce more individualized products cheaper and faster. “Customized just-in-time manufacturing will become inexpensive and common; we can all become designers,” she explains. “For instance, robotics can allow clothes in a particular style to be made to order from selected  fabrics; kitchen designs can be cut to order; packaged meals can be put together to order and in many cases delivered robotically.”

The impact on healthcare will be similarly huge. “Healthcare will be personalized and people will be advised in detail as to how to look after themselves,” Davies adds. “AI doctors will be able to diagnose, prescribe, and treat common conditions without consulting a human physician. Already, AI is monitoring glucose levels and suggesting doses of medication to diabetic patients with insulin pumps, and AI tests for other types of illnesses that are coming online.   Also, in the very near future, AI will be common in health centers to improve the diagnosis of MRIs, mammograms, x-rays, (and) cancers.”

Whit Andrews, vice president and distinguished analyst at Gartner, says automation is changing industries by allowing organizations to increase the scale of what they can accomplish. The technology also extends humans’ ability beyond their traditional limitations. “When you are effectively automating your processes, you’re now capable of doing more with fewer people, more with less resources, and it may open up new opportunities for you that previously have not existed,” Andrews says. “(When) you talk about any kind of artificial intelligence or robotic technology, what’s fun is you can do something in a tireless way. You have the same level of ability to do an evaluation the thousandth time as you do the first time.”

The industries that are addressing automation and AI most ardently are often the ones with the greatest amount of data, Andrews suggests, and this list includes financial services, as well as manufacturing, retail, healthcare, and higher education. Chris Roberts, associate director at Cambridge Consultants, says AI and autonomy in agritech is also an exciting realm right now.

“For me, the most exciting thing is that a lot of individual technologies are coming together to enable more tasks to be automated across many different industries,” Roberts says. “It’s one of the secrets of robotics—although robots have been in use for decades, most robots are actually pretty dumb. It’s only in the last few years as they’ve been combined with tech like machine vision and artificial intelligence that they’ve been able to access a whole new set of semi-structured tasks, where the objects they’re handling or the effects they need to achieve vary from time to time. In agritech, for example, farming has been automated for centuries, and while you get to combine harvesters for row crops, every whole apple you eat has still been picked by hand. Right now we’re seeing that starting to change.”

In agriculture, drones can help monitor large crop fields relatively cheaply, and robots with computer vision can help diagnose crop diseases or even pick crops like berries that traditionally needed a human’s touch. Cambridge Consultants has been working on projects in autonomy in agritech and fruit and vegetable handling, for instance, by designing Mamut, an autonomous robot that captures data about crop fields and individual plants, as well as an actuator that can replicate a human’s grip on a piece of fruit, thereby opening the door for robots to pick delicate produce without crushing or damaging it.

Source: McKinsey

Roberts says automation will drastically impact any and all industries with dangerous, dirty, or boring jobs, especially those struggling with a shortage of skilled workers. “Interestingly … all the clients who are working with us to develop automation solutions are doing so because they can’t recruit people to do the work, not to save money,” Roberts explains.

Source: McKinsey

Peggy’s Blog

Meeting Customer Expectations with AI

It’s truly a connected society and our customer expectations are rapidly changing. And as such, technologies are emerging and are raising our expectations and the bar for what customers are demanding by the day, and in many cases, by the hour.

Read More

Innovation Center Keeps Getting Better

How exciting is our industry right now? You might say I’m easily excited, and I’d probably agree, because I just love this stuff, especially when we are talking manufacturing and more.

Read More

AI on the Future of Work

As technology changes, so does the future of work and the future of workers. Therefore, it’s time to be forward thinking about the future of work and future workers and how industries will change as a result of AI (artificial intelligence) technologies, robotics, and even the  future of healthcare.

Read More

Future Worker Is Here 

The worker of the future is here. We hear it every day about the importance of focusing on the needs of the next generation and how we need to mentor them.

Read More

America’s Infrastructure Gridlock

From GM’s “See the USA in your Chevrolet” to “Get your kicks on Route 66,” American’s have expressed their yearning for the open road.

Read More

Industries with the biggest estimated dollar impacts as a result of AI,
based on 400 “use cases” that would allow them to use AI.

Retail:

$800 Billion

Consumer Packaged Goods

$486.3 Billion

Travel:

$479.7 Billion

Transport and
Logisitics:

$474.8 Billion

Automotive and
Assembly:

$404.9 Billion

Source: McKinsey

Socioeconomic Implications of AI

McKinsey’s analysis suggests the economic benefits of AI-based automation and innovation are secured at a cost, and that cost—described by the research firm as “negative externalities and transition costs”—could reduce the GDP by nine percentage points or around $7 trillion by 2030. AI technologies will render some skills obsolete. As the nature of work changes, the workplace will change alongside it, and the workforce will need to adapt. Despite this reality, humans still offer valuable qualities that AI can’t emulate—at least not yet.

Kris Hauser, associate professor in the Dept. of Computer Science at University of Illinois at Urbana-Champaign, says automation and AI are changing the game for the enterprise because any repetitive task involving human intelligence can, hypothetically, be done better and cheaper by a machine. “Moreover, machine learning algorithms can be better at integrating vast sources of information than humans,” he says. “Despite this great promise, the other side of the coin is that it is still difficult to responsibly engineer automated systems to avoid bias and to exhibit humanlike judgment in understanding context and ambiguity. Moreover, there are still fundamental technical challenges in certain tasks, like fine manipulation and navigating unstructured environments, where humans are still likely to outperform automated systems for decades to come.”

…Despite this reality, humans still offer valuable qualities that AI can’t emulate—at least not yet….

The industries that are most susceptible to automation, therefore, are those that involve repetitive, structured tasks, little judgment, and little direct interaction with humans. Hauser says in the near-term, the impact will be greatest in information processing jobs, like secretaries and cashiers, because physical machines are more expensive to design, build, and maintain than computers. But the impact of robotics on many industries will be enormous in the long-term.

What AI cannot readily do yet, according to the Brookings Institution (2019, pg 29):

1. Non-routine “abstract manual” activities (perception, manipulation, dexterity, physical adaptability)

2. Creative intelligence (ideation, critical thinking, problem solving)

3. Social intelligence (intuition, teamwork, persuasion, situational adaptability, perceptiveness, caring for others)

“The socioeconomic implications of autonomy and AI are vast, since the cost savings will make shareholders and consumers happy, while workers will be hit the hardest,” Hauser explains. “The socioeconomic benefits include greater profitability, more competitive manufacturing in the U.S., safer car travel, less traffic, cheaper goods for consumers, and household convenience. However, studies have predicted that these technologies will displace millions of jobs over the next decades in the U.S. alone.”

Hauser points to the reality that automation will not only affect blue-collar workers but also white-collar, information economy, and service jobs. “Entry-level information processing jobs are the most vulnerable to automation, whereas skilled trades like plumbers and electricians are some of the most secure. These trades require reliable judgment, creativity, and manual dexterity at a level far beyond robots’ current capabilities,” he says. “Our political systems are bound to be stressed by these changes, and as a society we will need to begin a deep conversation about how we cope with economic anxiety due to job losses, increased wealth inequality, and dehumanization of our communities.”

Wendy Ju, assistant professor at the Jacobs Technion-Cornell Institute at Cornell Tech, says it’s not clear what the net effect of the use of automation, AI, and robotic technologies will be.

“On one hand, these emerging technologies often lessen the number of people required to perform tasks; on the other hand, they are expensive enough that they tend to be employed where there just aren’t enough workers for whatever reason,” Ju says. “What is clear is that many traditional jobs will involve more robots and automation than in the past, much like most jobs involve computers now.”

In some ways, today’s regulatory agencies are not fully prepared to address the challenges of autonomy and AI. “This (is) such a new problem that the government does not quite know how to tackle it,” Ju adds. “I think a lot of the legislation to date has been rather ham-handed and ineffective. We need bright minds in information and law to tackle these issues.”

Kagan Tumer, director of the CoRIS (Collaborative Robotics and Intelligent Systems) Institute and professor in the School of Mechanical, Industrial, and Manufacturing Engineering at Oregon State University, says we should be asking what regulating AI even means. “AI is an umbrella term for algorithms that perform specific tasks like speech recognition, computer vision, sequence prediction, pattern classification, decisionmaking, game playing, and so on,” Tumer says.

“Sure, misuse of those algorithms can be harmful or create bias—(e.g.,) computer vision for face recognition, classification algorithms for credit assessment, prediction algorithms for parole decisions. But the problems stem from how the technology is used and what data was used to generate them, not the algorithm itself.”

Therefore, Tumer suggests the industry talk about laws and regulations for specific uses of AI algorithms. “For example, we can regulate how face recognition software is used in public places or airports. We can regulate how banks use predictive algorithms to do credit risk assessment. But it doesn’t make sense to talk about ‘regulating AI’ in general, because it’s not clear what that even means,” he adds.

Arizona State University’s Davies says the hard truth is that in an automated, AI-driven world, many people will never be able to find a job. How will we organize society when most people won’t go to work in the traditional sense? It’s a compelling question—one the industry must start thinking about now. Robots, cobots, automation, and other AI-related technologies are changing industries and society as a whole, and there are a lot of questions to consider as a result.

Estimated Impact of AI and Other Analytics by Industry, Function, and Business Problem.

Source: McKinsey

Transparency in AI will be crucial going forward. As Gartner’s Andrews puts it: People don’t want to talk to a monster if they think it’s a human. “The spooky part of the movie is where you find out that you’re talking to a ghost,” he says. “What you want is transparency. You want to know where you’re interacting with technology and where you’re not.”

AI machines and algorithms are expected to create 133 million new roles, 
but cause 75 million jobs to be displaced by 2022. 

(change in how many tasks between both years will be taken over by AI and automation)

Davies adds: “We can’t have confidence in the output of AI if we can’t trace the detailed workings of its algorithms. Apart from the possibility of errors in an algorithm, unconscious racial or gender biases could creep in, resulting from a coder’s past experience or selective training of the AI skewing the outcome. This could lead to unfairness, for instance, in candidate selection for jobs, access to credit, or an invitation to participate in medical trials.”

Understanding AI decisionmaking is an active area of research, and there are plenty of other questions waiting to be answered too. AI may reduce the prestige of human intelligence and creativity; the ramifications of this are unclear. AI will certainly rewrite the societal rulebook as we know it—in fact, it already is. Governments, industry, academia, and other entities are working toward ethical frameworks designed for a brave new world in which machines act autonomously. Already, these technologies have become powerful, and many believe regulation in some form may be imminent. What this will look like, however, is still up for debate.

One of the most interesting questions AI is raising as it revolutionizes life and business is this: What makes humans different from AI-enabled machines? And how can society harness the power of machines to enhance human life and not detract from it? “Futurists dream of ‘transhumanism,’ in which progressive cognitive advancement leads to a merger of human and machine intelligence, with AI forming a type of extended identity. The question ‘What does it mean to be human?’ is likely to change profoundly,” Davies muses. “Are we ready for this?”

Want to tweet about this article? Use hashtags #M2M #IoT #robotics #AI #artificialintelligence #automation #enterprise #robots #cobots #manufacturing #agriculture #retail #healthcare #data #machinelearning #autonomy

 

Transcription

Episode 628 | 08.20.19

Carlene C. Ileto:
Pivotal Moments in our Careers

Guest Contributors