What’s Next in 2017: Artificial Intelligence
I expected to feel out of place at the swanky university event for bioethics. My wife, a professor with expertise in the field, had invited me. But when I introduced myself as a software engineer to the attendees, many wanted to talk about one thing?—?artificial intelligence (AI): “How will it affect society?” “What are the goods and bads?”
Artificial Intelligence has been popping its head up everywhere. At a recent holiday event with extended family, a relative lectured my brother and I about how our companies?—?Google and Agolo?—?were destroying jobs by automating human tasks through AI.
Artificial intelligence has already entered the national conversation on a broad scale, but 2017 could be the year that we start seeing AI having more impact in our daily lives and causing disruptive change. But before we speculate on the potential downsides to AI, let’s consider the incredible benefits it can bring:
The U.S. Department of Transportation estimates that self-driving technologies could save up to 94% of the yearly 35,000 vehicle fatalities due to human error. That’s more lives saved in a year than the number of U.S. casualties in the Iraq and Afghanistan wars, as well as in domestic terrorism.
New consumer products that utilize AI promise to make us more efficient and to enable us to focus on non-mundane matters. Services such as a personal assistant or research assistant, which previously were only available to a minority, are now available for a small price due to AI-powered services (X.ai and Agolo are two examples).
Artificial intelligence may be able to shed light on problems that were previously unsolvable, such as cancer treatment and virus control, and can give us new insight into such integral processes such as education. An anecdote from the Go match between Google’s AI-powered AlphaGo and Fan Hui illustrates how AI can provide unique insight.
After playing a close match, AlphaGo played a move that initially shocked people in attendance. Many thought it was a mistake. But later it turned out that the move was exactly what turned the advantage to AlphaGo. “It’s not a human move,” said Fan Hui. “So beautiful.” (link: Wired)
In the game of Go, the machine is given a simple goal, for which it optimizes every interaction; win the match. But what about when artificial intelligence faces complex decisions and trade-offs? This brings us to some of the concerns over AI?—?that despite the complexity of the algorithms, engineers’ biases and short-sightedness can have powerful side-effects.
We can see several examples of this over the past year. Microsoft released an AI-powered chat bot on Twitter, “Tay”. However, within a few hours of the release, Twitter users had figured out how to get Tay to repeat Nazi propaganda and images. The engineers apparently hadn’t considered this “edge-case” and the bot was unprepared for the attack. (link: NYTimes)
Tesla released an autonomous driving technology in one of their vehicles, while insisting that it wasn’t intended to be fully self-driving. However, because the technology relied most heavily on light images, it failed to stop a vehicle that crashed into a white truck. Newer models are leveraging radar technology in addition to light images to reduce the possibility of error. (link: NYTimes)
Besides engineering mishaps, what about the human concerns? What about the jobs that artificial intelligence will replace? The recently acquired startup Otto wants to put self-driving freight trucks on the road. This could disrupt the livelihood of 1000’s of truck drivers. What will those people do?
Every technical disruption has a darker side. The industrial revolution in England improved textile manufacturing 50-fold, but it also created an oversupply of textile workers, causing many to lose their jobs. That’s actually where the term “luddite” comes from?—?Luddites were textile workers who revolted against the new technologies.
Many people are speculating on how to deal with this potential onslaught of automation. Modern Luddites would like to prevent autonomous technologies in order to protect jobs. Some technologists, like Elon Musk, suggest that governments may need to establish a base universal wage to support displaced workers.
Just like the complex problems faced by artificial intelligence, our own attitudes towards AI involve trade-offs. If we embrace technological progress and wish to continue as a leader in innovation, we also have to confront the potential loss of jobs and need for re-training. If we say no to automation and new technology, we risk dropping behind the rest of the world in technology and innovation. The trick will be getting AI to play for us, and not against us.
Like these insights, share this article
What do you see as the pros / cons of Artificial Intelligence?
This article is part of the LinkedIn Top Voices list, a collection of the must-read writers of the year.
A.M.S. Officer at Tanga Fresh Limited
7 年Proper balance between machinery and human race
Conceptual
7 年"a base universal wage" for me seems as adequate solution when capacity of work performed by AI will exceed the minimum, necessary to keep people alive. Let's see how it will work in practice. :)
AI/Deep Learning Scientist
7 年war robots programmers
Senior Architect at Health Catalyst
7 年As the saying goes, “AI won't replace workers, it will replace tasks.” And the fact that tasks are what is being replaced means that the direction of tasks will also become easier to automate. This entails several problems: First, it means that disruption will come sooner rather than later, since individual tasks can be replaced without achieving very high levels of general AI. But, fewer tasks means fewer workers, even if it is only sub-tasks being replaced. This is not new, it is similar to what happens when, for example, hamburger patties are shipped frozen or even pre-cooked and food service workers just have to finish the cooking process. The problem is, AI will accelerate this process beyond rate we've yet had to cope with. Second, the tasks being replaced will not cluster in any single portion of the work hierarchy. Many low-level tasks will be automated (think robot vacuum cleaners). However, many higher level tasks such as legal discovery and medical data analysis will also be automated. Third, management of tasks is one of the things that will become easier to automate, both as a primary effect of AI, and as a result of “task-thinking.” This will result in the acceleration of the use of workers as task commodities. It's true that this started long, long ago – even before the industrial revolution – but the pace will accelerate as will the level of detail in which tasks are assigned and monitored increases. This could result in a situation where humans function as essentially “dumb” appendages to the machine, doing piecemeal tasks simply because we haven't gotten around to automating them. (This scenario is covered in the first half of Marshall Brain's book “Manna.”) What's the solution? I think the real problem is not automation, it's one of treating human beings with dignity and making sure that all can share in the benefits of automation. Essentially, it's a political and social problem.
docente en ministerio de educacion boliviana
7 年Yo hablo espa?ol ups