Artificial intelligence (AI) isn’t business as usual.
Because of complementary advances in natural language processing, machine learning, and image recognition, the range of tasks for which AI is well-suited is growing daily. And when a critical level of AI saturation is reached, we anticipate profound disruption in the world of work.
Imagine a barrel perched on one end of a seesaw; it represents the capacity of AI. On the other side of the fulcrum sits a giant of a man, a stand-in for human labour. A garden hose runs into the barrel, slowly filling it with water. As the person watches, confident at first in his weight, his opposite party gets heavier and heavier until a critical mass is reached, tipping the scale and shifting the human, perhaps slowly at first, but eventually, with real force and conviction. The giant, once confident, is now definitively the lighter of the two.
Once that scale is tipped, once the capacity of AI outweighs the capacity of human workers, there’s no going back or shifting weight onto the flesh and blood side of this balance.
Bank of America, for instance, a company not given to wild projections, estimates that within 30 years, advances in artificial intelligence will cost 230 million knowledge workers their jobs. In the US alone, they suggest that as many as 47% of employees will cede their place to machines.
This may sound alarmist, but our research supports this estimate. To understand the scope and scale of the coming change, you’ll need to understand the basics of the tech driving disruption.
Natural language processing
Natural language processing (NLP) is the ability of a computer to recognise and interpret ordinary speech. It enables conversational computing, allowing users to interact with machines much as they would a person. If you’ve used Apple’s Siri, for instance, or Amazon’s Echo, you have a sense of rudimentary NLP. But the state-of-the-art is miles ahead of these; remember, IBM’s Watson was able to defeat two champions on Jeopardy, a feat that required the machine to understand word play, innuendo, and humour.
And make no mistake, natural language processing isn’t all fun and games.
Intelligent chatbots are being embraced by industry and government alike for their ability to provide interactive, immediate customer solutions. Habituated to instant replies by texting and social media, neither B2B or B2C customers want to wait. Local governments in the US are exploring chatbots for the DMV in an effort to reduce congestion and streamline workflow. And in IT, for instance, the overwhelming majority of calls relate to forgotten or incorrect passwords. Intelligent chatbots are an ideal solution to this customer service nightmare: they’re available day and night, responding instantly with a friendly, patient password reset. This frees IT specialist to tackle the real work, the true problems, while keeping more quotidian customers happy.
Machine learning and image recognition
Machine learning is the process through which an artificial intelligence teaches itself. Making use of layered processing, the most successful forms of machine learning mimic the function of the human mind. Data is collected and analysed at lower levels, where patterns are recognised and passed to higher levels for further ‘thought.’ At each stage, the gathered information is subject to scrutiny by ‘smart’ algorithms, until, at the end of this only milliseconds-long process, an answer to a question is produced with an accompanying sense of its probability of error.
For instance, if an artificial intelligence is tasked with facial recognition, it will scan images looking for basic facial features: the face itself, and strong ‘landmarks’ that help it ‘see’ the difference between a person and a potted plant. At the lowest levels, its machine learning is comparing what it’s seeing now to what it has learned in the past, refining the process; this previous experience is everything. The more and better the inputs, the better the algorithms will function.
In comparison with these thousands of previous images and self-tutorials, the machine intelligence identifies what it believes to be a face and passes this information ‘up’ for further analysis. An eye is identified, and then a nose and a mouth. Precise measurements are made, the face is mapped, and then, near the end of the process, the machine compares this facial map to known images, adjusting for differences in lighting, partial views, and a turned or lowered head. In the end, it identifies a suspected terrorist with 99.8% confidence.
This gives you a sense of the power of machine learning and image recognition, and it’s already powering everything from radiology (Enlitic) to beekeeping (BeeScanning). Indeed, the range of applications of this adaptable tech is startling, and the buzz about AI is more than mere hype. Forbes reports that 38% of companies are already leveraging the power of artificial intelligence to redefine business, a number set to nearly double by next year. And Forrester Research thinks that investment in AI will grow by as much as 300% this year alone.
What does this mean for you?
As AI gets smarter, cheaper, and more broadly applicable–a time-frame better measured in months than years–it will slowly replace human beings in any repetitive, predictable role. That may not sound revolutionary, but ‘repetitive and predictable’ covers a lot more ground than you might expect, running the gauntlet from medical diagnostics and surgery to the practice of law and legal research to even engineering and writing. Watson, for instance, has already directed a film trailer, earned a position at the Sloan Kettering Memorial Cancer Center, and been hired by Bon Appetite to develop recipes. That’s just the first generation of real AI.
In general, it’s better to ask, “What can’t AI do?” than to doubt its capacity.
Indeed, a sense of this broad applicability lead the World Economic Forum to predict a “Fourth Industrial Revolution” that will transform the way the world works–indeed, it may determine whether the world works at all in the traditional sense. As machines adopt routine work of all kinds, both blue and white collar, how will we adapt to the change?
Ginni Rometty, the CEO of IBM, promises that thinking machines like Watson will augment, but not ultimately replace, people at work. In her view, the real value of AI is that it can bear the burdens of value-added employees, freeing them for tasks that demands the human touch. But futurists like Faith Popcorn aren’t reassured, and even Elon Musk has begun to publicly worry about the future of flesh and blood in the era of silicon and steel. If Watson can write for a culinary magazine and exceed the capacity of human physicians to diagnose lung cancer, is there really a role for people at work?
Only time will tell.
But this question need not be something to fear. Marcel Bullinga, a futurist in the Netherlands, reminds us that AI has the capacity to revolutionise our daily lives. “AI will conquer the world, like the Internet and the mobile phone once did,” he thinks. As it does, and to the degree that we can find the strengths of people, magnifying them with the heavy lifting of machine intelligence, we can reimagine that seesaw as a chance to elevate human labour, amplify its effect, and carry it to new heights.
For more see:
- Yes We Can Have A People-Centered Internet (PCI): Join the PCI Digital Future Event 10 Dec in San Jose
- Amazon opens its internal machine learning courses to all for free
- DISA’s CIO Dave Bennett on Achieving Efficiencies
- Doing Business with DoD and All of Fed Government- Post-Sequestration
- Interesting Take: Forget Watson, the Red Hat acquisition may be the thing that saves IBM