Developer versus the Bot. Or… featuring the Bot?

16 June 2017Strategy

Things are changing quickly in the field of digital marketing and technological innovation. The big players are increasing their investments in artificial intelligence (AI) and AI-related startups. Many recent articles describe the fear of AI taking over human jobs while leaving us feeling useless. Is this imposed fear valid? Previously, we explored what AI means for content creators and designers. Today, we’ll have a look at what AI means for the future of developers.

These days, consumers are quite aware of what machine learning can do for them. But do they realize what it takes to make machine learning possible in the first place? We’re sure a lot of consumers (and even digital marketers) are not familiar with the complexity and developer effort behind machine learning applications. Could we be overstating the impact of AI?

Let’s do a reality check to see if and how fast AI is becoming a threat to developers. For this exploration, we’ll take into account The 7 Stages of Robot Replacement as proposed by Kevin Kelly:

  1. “A robot/computer cannot possibly do the tasks I do.”
  2. [Later] “OK, it can do a lot of those tasks, but it can’t do everything I do.”
  3. [Later] “OK, it can do everything I do, except it needs me when it breaks down, which is often.”
  4. [Later] “OK, it operates flawlessly on routine stuff, but I need to train it for new tasks.”
  5. [Later] “OK, OK, it can have my old boring job, because it’s obvious that was not a job that humans were meant to do.”
  6. [Later] “Wow, now that robots are doing my old job, my new job is much more interesting and pays more!”
  7. [Later] “I am so glad a robot/computer cannot possibly do what I do now.” [Repeat]

How fast is AI catching up?

To get an indication of the current state of artificial intelligence, let’s have a look at one of the frontrunners in AI technology: IBM. What are they up to?

IBM is currently using their Watson platform to venture into oncology, and it’s making quite a bit of waves. Fortune explains:

“The platform was fed textbooks, medical journals, and possible treatments and then trained on cancer causes. As a result, Watson is now able, in some instances, to spot cancer better than a panel of human experts—an outcome that points to AI’s potential to find solutions to the world’s most unsolvable problems.”

Surely, that’s great news. But it also shows signs of a future where computers will no longer need human intervention to solve our biggest issues. Will robots take control? Earlier this year, IBM CEO Ginni Rometty downplayed the fears of a robot takeover, claiming that AI will only augment human work, not replace it. Will that augmentation apply to all fields, though? What about the work that developers do? Can it be replaced by bots?

AI sucks at coding just as much as you do

Artificial intelligence that knows how to write its own code is already here. In cooperation with Cambridge University, Microsoft has created an AI that can solve programming problems by taking existing lines of code from other sources. The program is called DeepCoder. When you give it a certain input and the desired output, it will automatically find code that will transform that input into the output you’ve requested. It can do so in a matter of seconds.

(Image by Matt Q Spangler)

(Image by Matt Q Spangler)

Basically, all the AI does is reuse existing code snippets to see if they generate the desired results. This is why you could argue that Microsoft’s DeepCoder sucks at coding as much as the typical Stack Overflow user. After all, the bot is just using copy/paste without actually knowing what’s happening. Still, it’s a promising approach. Right now, the program can solve problems that take five lines of code, but its ability to use machine learning will soon make it able to tackle more complex challenges.

Robots will teach other robots

In the future, knowing how to code may not even be necessary in order to teach AI new skills. Scientists are currently exploring how to make robots learn new movements without having to program them.

The idea is to let a robot watch a video of another robot doing a particular movement. The robot will then try to mimic this movement by combining the right, elementary moves from its movement library. In a way, this is similar to how a child learns to make complex movements: it breaks the movement down into shorter, less complex movements that it already knows. And just like humans can transfer their knowledge to their peers, robots will soon be able to transfer their skills to other robots.

So, AI is learning how to code and robots can learn without being programmed… Where does that leave us and our jobs?

Robots will replace 50 percent of all jobs

Kai Fu Lee

Kai Fu Lee, futurist and founder of venture capital firm Sinovation Ventures, predicts that robots will likely take half of all human jobs over the next decade. According to Lee, artificial intelligence is something that will be “larger than all of human tech revolutions added together, including electricity, the industrial revolution, internet, mobile internet.” Intelligent decision-making engines are what will replace humans in their jobs, says Lee.

Lee does make an exception for jobs that are centered around human interaction: “Touching one’s heart with your heart is something that machines, I believe, will never be good at.”

What is the half time score?

While it’s true that some aspects of human routines will be automated, we believe this will ultimately be beneficial to developers: it will give them the opportunity to focus on other (more complex and challenging) tasks. That’s a situation similar to the one we found for designers.

After considering the examples above, we can conclude that we’ll see ‘Developer featuring the Bot’ rather than a situation where the two will battle each other. Without developers, there would be no bots in the first place. Bots still cannot learn completely by themselves, so humans will have to ‘raise’ and teach them by providing the right data to these systems. Developers will need to build these systems and think out the ‘right’ algorithms for them to work with.

As McKinsey research explains, 49% of time spent on work activities can be automated. But replacing development entirely? No. Not anytime soon. That means that, at least for now, developers are still at stage 2: “A robot can do a lot of my tasks, but it can’t do everything I do.”

Interested in how AI will affect other expertises? Will your tasks be replaced in the blink of an eye, or will it not go that fast? Let us present your future. Contact ai@afrogleap.com.