Artificial intelligence (AI) doesn’t get cranky, or spiteful or distracted by fantasy football. What AI lacks as a stimulating conversation partner it makes up for in diligence and focus—and that’s why so many organizations are looking at ways to build AI into their everyday workflow.
What does that mean for workers, for executives and for the workplace of the future? The answers are as diverse as the different types and shapes of AI.
“We keep talking about AI like it’s singular, like there’s an AI, but the reality is that there are lots of pieces of AI in lots of places,” said Genevieve Bell, an Intel senior fellow and director of the Autonomy, Agency and Assurance (3A) Institute.
Some AIs will continue to quietly work in the background on problems like fraud detection and face recognition, indistinguishable from any other kind of computer automation. Others will surely expand on the growth of voice interfaces to act as assistants, agents, possibly even supervisors.
Everyone has a distinct set of expectations and apprehensions about how AI will impact the workplace, but here are five things every worker should know today.
1. AI may be around the workplace before it’s in the workplace.
AI’s importance to smart building and smart grid projects means that your first workplace encounter might be in the infrastructure of your workplace, rather than on the job itself. For example, AI-guided elevators follow complex rules that prioritize stops on floors with high demand and make predictions about rider volume, rather than following the typical method of simply stopping in order at every requested floor.
Smart elevators are a proxy for the overall impact AI is likely to have. In time, people will learn to change their expectations about what an elevator ride will be. Meanwhile, engineers will improve the human-AI interface so that elevators can better communicate expected stops and potential disruptions to concerned riders. Think of these daily encounters with AI elevators as a trial run for much greater exposure, while keeping in mind that no two processes will be disrupted in quite the same way by AI.
“The way lifts [elevators] become artificially intelligent will be different from power grids, or traffic lights, or something new your phone can do,” Bell explained. “And how we engage with it, how we feel about it and whether we even notice it is going to be different not only depending on the application but also the sector, the object and the relationship we already have with it.”
2. Tomorrow’s AI won’t look (or work) like today’s.
Some AI innovators have aggressively marketed their solutions—with many becoming household names—by being available on hundreds of millions of smartphones and other consumer devices. But in much the same way as office technology has radically changed over just a few short decades, AI is going to leap well beyond its current form.
Some of what is described as workplace AI, such as advanced analytical tools, decision-making aides and process automation, isn’t really capable of learning or improving its own algorithms.
“Any one of those is interesting, and most workers are encountering those already in various forms,” Bell said. “But none quite represents AI.”
Just as cumbersome, nonergonomic computer terminals gave rise to carefully engineered interfaces, much of what changes about AI in the workplace will come from experience and research. The early generations of AI are more focused on function than form, but that dynamic will shift in time.
“Machines need to know more about how humans behave, and humans need to get better at thinking about what the right relationship with machines is,” Bell noted.
3. AI can’t (and won’t) do everything.
A word about AI’s inherent weaknesses: It’s not easy to design AI to be ambitious or inventive. That’s why AI has shown much more potential in rote tasks such as analyzing contracts and helping people deal with parking tickets than it has in more creative pursuits.
“Pieces of jobs that are data-rich and rule-heavy are likely to see more AI augmentation more quickly, in areas like law, medicine and accounting,” Bell said. “Just like early waves of automation affected human labor where there were repetitive [manual] tasks, these are repetitive cognitive tasks.”
“Machines need to know more about how humans behave, and humans need to get better at thinking about what the right relationship with machines is."Accordingly, business leaders need to think about engaging AI on processes where it can make a tangible difference, and where it is both possible and worthwhile to dig into the gritty details to build rich data sets and reliable algorithms. Some seemingly repetitive and rote processes are actually laden with hidden intricacies, workarounds and undocumented procedures that are difficult to capture and relate to AI.
“Lots of things are done in informal ways, so when you try to automate the process, everything suddenly stops working,” Bell said.
In fact, the leaders who do the best with AI may be those who help define new metrics for AI success. Pursuing the same old goals of the previous millennium may not be enough.
“The Industrial Revolution proceeded on the basis that efficiency and productivity are the right things,” Bell said. “I’m not sure that’s what this next generation of technologies needs.”
4. Flexibility is key to adapting.
If you’re concerned about your job being disrupted or automated away by AI, one crucial thing you can do is to get better at things humans already have an advantage in—but where machines struggle.
“The skills that become valued are flexibility, nimbleness and how you handle both ambiguity and uncertainty,” Bell said.
It will also require flexibility to get accustomed to a world with fewer gray areas. When AI is charged with diligently enforcing widely agreed-upon standards fairly and transparently, it becomes difficult to break those patterns and achieve a more fluid outcome. For example, AI designed to enforce traffic regulations is unlikely to ever let a speeding driver off with a warning.
“Humans are used to being able to negotiate around the edges,” Bell pointed out. “Some of that is going to go away, and I don’t think we’re quite ready for that.”
“The skills that become valued are flexibility, nimbleness and how you handle both ambiguity and uncertainty.”5. Our ability to ask good questions may be the most valuable skill of all.
AI can collect and process data and present a course of action in a matter of microseconds, far faster than the reaction time of any human. In this context, people will thrive—as both collaborators with and checks on AI—by being better at asking the crucial questions: Where are data and recommendations coming from? How was a decision reached? Is a given recommendation actually the right thing to do?
“We need to talk about critical thinking, and how to move from thinking about solving problems to learning how to frame the right questions,” Bell said. “We need to teach primary and high school students that rules aren’t the answer to everything.”
Beyond the scope of these five points are countless more questions about how AI will reinvent work as we know it. It’s OK to keep asking those questions. In fact, it’s the most human thing you can do.
“One of the main sources of confusion is still how to talk about all this,” Bell said. “But the opportunities are out there for people who know how to ask hard questions.”
About the Author
Jason Compton is a writer and reporter with extensive experience in enterprise tech. He is the former executive editor of CRM Magazine.
CREDITS: SolStock/Getty
No comments:
Post a Comment