New book explores how AI is really changing the way we work
One of the earliest predictions about artificial intelligence was that robots would steal people’s jobs, especially for seemingly monotonous and menial tasks.
Example: Flippy 2, the “autonomous robotic kitchen assistant” developed by Miso Robotics to flip burgers as they cook. The CaliBurger restaurant chain is testing Flippy at different locations. But what the robot does best is not what it was designed to do.
“[A CaliBurger franchise owner] said, “Flippy isn’t as good as I’d like to flip burgers, but it’s pretty good at popping baskets of fries out when they’re done,” said Thomas Davenport, a digital researcher at the MIT Initiative on the ‘Digital Economy. “We’ve known for some time that AI and robots could take the place of burger pinball machines, but that hasn’t happened yet.”
Davenport’s new book, “Working With AI: Real Stories of Human-Machine Collaboration”,” examines 29 examples of humans working with AI-enabled systems, showing when and how AI works best and how companies can use AI responsibly. The book is co-authored by Steven M. Miller, a professor at Singapore Management University.
The common theme is that AI augments the work that humans do, without fully automating it. That means many jobs are secure, Davenport said during a recent webinar hosted by the MIT Sloan Management Review.
“One of our main conclusions is [augmentation] does not appear to result in job loss,” Davenport said.
Davenport and Miller explained how companies best implement AI:
AI works best under “normal” conditions
Miller defined automation as a process in which a machine performs enough tasks that the human who previously performed them is effectively displaced. Augmentation, on the other hand, is the automation of smaller tasks, not a full job.
“The human remains involved in the direct execution,” Miller said.
This distinction is important and helps explain the scenarios in which organizations would seek to use one or the other.
“AI is great at running a machine under what we consider normal conditions. Under the expected circumstances, AI is really good at getting things done,” Davenport said.
An example is searching for keywords in long documents such as legal contracts or medical records and producing results in seconds. Another is to monitor information from thousands or even millions of sources, whether it’s sensors at a manufacturing site or cybersecurity monitoring tools. These types of processes can be labor intensive for humans, but easy for machines.
“Full automation works great when you’re in familiar territory and when the nature of the variance and the nature of the unexpected are somewhat familiar. You can really aim for that efficiency, consistency, optimization, and productivity,” Miller said.
Of course, conditions are rarely normal enough for AI to operate without human intervention. That’s why Flippy needs an experienced cook to figure out which part of the grill is the hot spot that cooks burgers faster. That’s why the robotic weed puller developed by agricultural company FarmWise always needs a farmer walking behind it to make sure the plant it’s grabbing isn’t a vegetable. (Although it still beats the historic technique of spraying the entire crop with dangerous chemicals to kill weeds.)
That’s why self-driving vehicles work well in the dry air of Las Vegas or Phoenix, but haven’t made their way to rainy or snowy climates.
“I dread the day they come to Boston in the middle of winter,” Davenport said.
4 situations where boosting AI makes sense
The concept of augmentation is not new. Humans started using stone tools thousands of years ago. It is also familiar to knowledge workers who no longer need a calculator on their desk.
“Anyone who has used a spreadsheet understands the power of augmentation,” Miller said.
Yet the augmentation enabled by AI is often underestimated in today’s businesses.
“There are more opportunities than people realize,” Miller said.
Davenport and Miller provided examples that fit into four broad, overlapping categories:
- When companies want to experiment. Product development is particularly expensive in the pharmaceutical industry, where the median R&D cost for new drugs is now over $1 billion. Instead of spending up to a year working on a drug only to realize it won’t work, pharma makers are building an AI infrastructure to evaluate their data and assess potential use cases for a drug. new drug in as little as two weeks, Davenport said. Not only does this support frontline workers, but it also leads to new “data product” roles to handle these AI use cases.
- When there is a lot of pencil-pushing. “Every business person we meet, and every company we talk to, almost always has the same complaint: ‘We have so much to do, and we don’t have the manpower to do it.’ At the same time, they complain about the amount of heavy work they have to do,” Miller said. In these situations, he added, automation will free up employees to do more work. AI-augmented experiments, from product development to predictive analytics.
- When things go fast. Cybersecurity monitoring software can assess threats from millions of sources in real time. Some tools can also prioritize threats and even respond to them automatically, shutting down a compromised device, for example. But in “complicated and nuanced cases,” Miller said it takes an experienced security analyst to assess all the data collected and determine whether it’s a one-time incident or an indication of a pattern. attacks that may require larger scale mitigation.
- When workers want a recommendation. When consumers trust healthcare professionals, the increase acts as a form of decision support. For example, the AI looks at the available information and makes a recommendation, but the final decision is up to the doctor at the computer. In countries like China and Indonesia, where hundreds of millions of patients have access to AI-powered diagnostic tools like Ping An Good Doctor, “A doctor actually needs to produce the diagnosis and the treatment, otherwise it’s not valid,” Davenport said. “But the AI system saves the doctor a lot of time and effort.”
How to work with AI responsibly
Moving forward with the rise of AI means accepting that machines are better than humans at certain tasks.
Again, this concept is not new. “
We’ve always had great tools. We have jets, we have cars, we have excavators and cranes,” Miller said. “Now we have great tools that can do non-routine, cognition-driven things.”
At the same time, machines cannot integrate all dimensions of intelligence. Citing the adaptive intelligence theory of Robert J. Sternberg of Cornell University, Miller said the human mind is uniquely suited to creativity, wisdom, and context — not just realizing that something can be done, but whether it needs to be done. In other words, don’t underestimate what humans can do.
The most effective and responsible use cases for AI augmentation are not limited to rudimentary tasks or task automation. The goal should be to bring humans and machines together to achieve a higher level of intelligence than a human or a machine could achieve on their own, Miller said. The whole must be greater than the sum of its parts.
In doing so, companies can expect to create new types of products and services, as well as new types of work. But that comes with two clear caveats: workers need the right skills to use AI systems, and they need to see that AI is helping them and not going to replace them.
“We really need to step up and help people adjust to some of these new tools,” Davenport said. “We need to retrain workers and engage them in this new kind of work.”
Watch the webinar
Read next: Sec’s Gary Gensler on how AI is changing finance