You are here: News

Latest News

©iStock.com/Mustafa Hacalaki

June 02, 2014 by

Article Series: How Far Up the Ladder Can Automation Go? (Part 3)

This is the third article in our series on the positives and negatives of automation. As a manufacturer of automation machinery, we feel that it is our responsibility to be forthright about the controversies and questions that are a part of the industry. We provide these articles in an attempt to have a more informed customer and partner base, and in so doing, we hope to encourage a healthy dialogue on many of these topics. Rachel Greenberg writes marketing and technical content for Automation GT.

In 1964 episode of The Twilight Zone, “The Brain Center at Whipple’s,” an unscrupulous manufacturer replaces his staff with automation until finally he himself is replaced with Robby the Robot (no, really, Robby the Robot). This is a pretty typical example of a technophobic response to automation, and this episode captures a general feeling of apprehension about automation that was becoming more prevalent in the 1905s and 60s as automation in the factory was taking off. But in truth, we have now gone well past the point of automating to the level of management.

Consider the termite robots that made news a few months ago: these robots are able to coordinate on large tasks with a hive mind in imitation of the way termites coordinate. This eliminates the need for management as these robots are able to monitor themselves and correct for mistakes internally, as a group.

So we can automate rote manual tasks, and we can automate the coordination of rote manual tasks, but when performed well, management, like many jobs, involves a lot of creativity and clever problem solving, even when the employees are now robots. As long as we have human customers, management is likely to face many very human problems. Yet we are making advancements in automating many jobs that have historically been considered creative and high-intelligence jobs.

For example, a Los Angeles-based journalist recently made the news, not for the news he reported, but for the way he reported it. His algorithm, Quakebot, generates an article instantly whenever there is an earthquake in LA at or above a certain magnitude, thus allowing the newspaper to be the first to report on the earthquake without the journalist having to get up in the early hours of the morning to research the event.

In this case, as in other cases where portions of the journalistic process have been automated, the automation process supplements the job of the journalist. It does not replace it. The algorithm generates a simple and generic article that gives the basic facts. However, journalism falls into the category of what is considered higher-level employment because the act of reporting is an intellectual activity and is generally not thought of as dull, dangerous, or dirty (though in reality much of reporting may well fit into one or more of these categorizations).

However, the benefits to automating journalism are similar to the benefits in automating manufacturing processes: articles will be more consistent in their quality and more factually accurate, reporting will be less subject to bias, and articles will always arrive by their deadline and with no wasted time. The programs are cheap to use after they have been developed, and they will “never ask for a sick day.”

But still, if you go to read about Quakebot, you will come across such ominous titles as:

news4news3news2news1

 

 

 

 

 

It seems that whenever a job is automated, someone will use the some version of the phrase “a threat to professionals”: a threat to journalists, a threat to managers, etc. In this case, as per usual, the inventor of Quakebot insists that the algorithm is just meant to help journalists, not replace them. But the bottom line is that by using Quakebot, this newspaper does not need as many entry-level journalists who get the unpleasant assignments that involve getting up in the middle of the night to write these kinds of articles. In theory, every journalist would have one of these robots to help increase efficiency, but in practice, there is not an endless supply of news to report—the productivity of automated journalism is finite and can produce so much work that there is literally nothing left for an entry-level journalist to do.

There has already been lots of speculation about what automation means for people in search of entry-level careers as these are usually the jobs that are first to be automated. With fewer entry-level jobs available, more young people will be unable to get work and so will be unable to gain the skills they will need to advance their careers, thereby cauterizing their careers at the root.

In truth, however, it is difficult to determine exactly how much of a role automation might play in the disappearance of entry-level positions. As with the issue of job loss that began in 2000, it is nearly impossible to label a single cause. The general loss of jobs and market stability in America is a far more evident cause for loss of entry-level jobs, as a lower job creation rate across the board will include a lower rate of creation for entry-level jobs. As companies have cut costs and positions, entry-level spots are often the first to disappear because the young people who require more training do not produce as high a return on investment right from the hire date the way more experienced employees do.

However, because automation of high-intelligence jobs is still rare, we will, for the time being, likely only see it used as a supplement and not as a replacement. But even if we can get to the point where we can totally automate many entry-level types of jobs, what about the functions that require still more creativity? For example, as above stated, automated journalism decreases the likelihood of bias and opinion getting into an article. But what about the situations in which we want opinion, in which opinion improves our experience of a piece? As an example, consider the career breakout piece of Geraldo Rivera, Willowbrook: The Last Great Disgrace. This exposé on the deeply corrupt practices of this children’s mental hospital required a certain bias in order to be successful. It required emotionally charged language to produce the viewer response that Rivera wanted, the response that ultimately led to the shutdown of Willowbrook and earned Rivera a Peabody Award.

Of course, it wasn’t so long ago that the automation of a job like basic article writing seemed like science fiction. And there are some amazing ways that engineers have automated jobs that seem completely out of reach of automation. For example, there are now robots that can perform such tasks as:

  • Composing music: Emily Howell is a program that has composed music for piano, which, its creator likes to brag, has fooled scholars who have been unable to tell that it was written by a robot. Though Emily’s creator, David Cope, has made enemies in the past for his insistence that there is no real difference between Mozart and a computer, and that it isn’t soul that makes music, but rather careful mathematical nuance, in fact, Emily is now nearly outdated. There have now been scores of robots that write their own music, including one robot that can jam.
  • Creating original paintings: The Painting Fool is an AI program that creates its own unique paintings, sculpture, and multimedia projects. It can analyze human emotional response, recognize recurring images as parts of themes (for example, can recognize different types of flowers as part of the group of flowers or that images of both skyscrapers and street lights belong in the group of things included in a city scape), and can create abstractions based on visual input.
  • Home aids and nurses: ERWIN, the Emotional Robot with Intelligence Network, is currently under development to become the most socially aware robot on the market. Researchers have been working with the robot to determine what all goes into the perfect robot social behavior, what amount of facial expression, body language, and responsiveness make humans feel most natural in their interactions with the robot. This will then be factored into the design of a robot that can be used for home aid and nursing work for the elderly, people with disabilities, and children with Autism spectrum disorders.
  • Performing surgery: Though the fully autonomous surgical robot is unsurprisingly proving elusive, major advancements have already been made in automating certain parts of surgery such that robots can act as assistants to surgeons during procedures. Robots have superior ranges of motion and can handle highly finessed movements, and so when combined with the ability of a surgeon to make decisions and recognize complications, autonomous surgical robots can greatly improve efficiency of procedures and can cut down on invasiveness.
  • Composing poetry: Try your hand at Bot or Not. This website quizzes you on your ability to differentiate between poems that were written by people and poems that were written by programs. What do you think: Bot or Not?

poem1

poem2

 

 

 

 

 

 

 

 

The first poem is by a human, the second by a robot.

In her very compelling article on Bot or Not, Johannah King-Slutzky notes that the biggest give-away that a poem was written by a human and not by a robot is that humans will sacrifice consistency of form in favor of insight and meaning. But that doesn’t mean that we can’t make a program that can’t factor in “insight,” that can prioritize meaning over form. It simply means that we have identified the next thing to perfect. And it certainly doesn’t mean that a poet never wrote a poem that that did not prioritize insight over meaning.

So does this reiterate the question of David Cope as mentioned above? Are higher-level intelligence skills just equations that we haven’t quite programed in a sufficiently nuanced way yet? Could there potentially be no difference in quality between human and automated work, or even further, could automated work surpass the quality of human work in all fields as it does in dull, dirty, and dangerous jobs?

Part 1: Why automate in the first place?

Part 2: Who’s afraid of a little automation?

Part 3: How far up the ladder can automation go?

Part 4: How do we keep people employed? Or shouldn’t we?

Part 5: Addressing the concerns of the displaced

Part 6: How can we automate responsibly?