Make Innovation Everybody's Business

Could AI replace the Innovation Manager?

With the introduction of Artificial Intelligence, machines are moving up the career ladder from “blue collar” to “white collar”.

We wonder if “black turtleneck”, as in creative leaders like Steve Jobs, are next? Nosco Innovation Analysis, looks at events and trends through the lens of innovation. In this post, we’ll look at AI.

AI has been making headlines lately : In May 2016, it was reported that Foxconn, the iPhone manufacturer, were replacing 60,000 workers with robots. That’s more than half their workforce. Personally, I can report that last time I called Microsoft, I was 2 minutes into the conversation before realizing that I was, in fact, having a conversation with a computer

Also in May of this year, a major US law firm introduced its latest employee, “ROSS”. Dubbed by the company as “The world’s first artificially intelligent attorney”, ROSS can sift through thousands of cases in seconds. It then presents its findings to human lawyers, who can upvote and downvote the suggested findings. That way, ROSS is constantly learning, getting better and better with the years. Who knows? ROSS might end up making partner in firm. He can certainly put in the hours.

They’re probably used to it, but those in military service shouldn’t feel safe either.  In June this year is was reported that  “ALPHA”, an AI for military drones had gotten to a level where even the most astute tactical human experts were unable to score as much as a single “kill” in the simulator against it. According to the report, when handicapped with a slower drone, less weaponry and worse maneuverability, ALPHA remained utterly unbeatable and devastatingly deadly at the same time.

As we see, computers, robots and machine learning are not just replacing manual labour, they’re also doing service jobs and knowledge jobs. The more mechanical our jobs have been, it has long been a risk that machines will take our jobs. With the introduction of AI, machines are gradually moving up the scale from blue collar to white collar. Could the “black turtleneck” be next? Because AI is getting creative.

Artificial Creativity

More headlines: June 2016. The sci-fi short film “Sunspring” has been making the rounds on the web. In essence, the script makes zero or little sense. The writer in this case is a neural network for language recognition, a type of artificial intelligence. This AI named itself “BENJAMIN”.

BENJAMIN” was fed with a large number of sci-fi movie scripts, and out popped the script for Sunspring. It was then recorded by real life human actors. The AI also wrote the title track from the movie, after being fed the lyrics of 30,000 pop songs. Again, it’s mostly nonsensical. I’m sure you can think of a human-written track or two, where the same applies.

And there’s more happening in the world of music. Google’s “Project Magenta” is an AI that generates art and music. It “released” its first track this summer. Again, it’s hardly something you’ll enjoy.

AI in the innovation department

This brings us to this newsletter’s main question: What could this mean for the innovation department? For companies that work with ideas, and lots of them, this could be big. Nosco’s goal is to turn companies into networks of collaborative idea workers, so naturally we are interested in this too.

We’ve looked into some areas of collaborative innovation where AI could play a part by assisting the innovation manager and maybe one day replacing some of the jobs carried out in innovation. Rest assured there are also key aspects of innovation that we believe requires a human touch.

AI  for idea generation

An area of idea generation where AI could prove useful in the present state is what we call synthetic creativity. This discipline is about recombining and rearranging well-known concepts into new ones. In the world of startups and innovation this is often seen. Entrepreneurs and enterprises are striving to create the “Uber for X”. Hypothetically, the innovation department could set up an innovation AI, feed it some parameters for what ideas it would need, and have the AI churn out ideas. Hundreds of them to begin with. Thousands, if not millions over time.

AI for idea selection

As our AI and human powered innovation department  outputs ideas, we could have domain experts rate them and offer feedback. Much as we saw with ROSS, the lawyer-bot, an innovation AI could learn from the up-votes and down-votes provided. Our innovation AI could instant prepare shortlists out of thousands of ideas, for human evaluators to select from.

AI for testing ideas

An AI for innovation could have access to an abundance of data. Companies like Microsoft (Especially after the LinkedIN purchase), Facebook, Amazon and Google have so much data, connected in so many patterns about our behavior, they probably know us better than we know ourselves. Given access to this Big Data, AI could simulate massive, instant, focus groups. This could allow the innovation department to test thousands of permutations of their offerings in minutes.

Only for incremental innovations?

Machine learning, the basis of AI, is based on data. In other words, datasets from previous events.  One could imagine that innovation AI would be almost perfect for incremental innovation. Solving such tasks as finding the “sweet spot” for your offering, in terms of features and price, would be right up its alley. But what about the radical innovations? Could predictions on these show up as weak ideas to the AI? Could we end up discarding fantastic innovations on account of this?

What about the human touch?

In our practical experiences with idea platforms, there is an undeniable element of the human touch. We see this is two major ways. One is “hunches” and the other is “cultural fit”.

Both of these bear great influence on which ideas are selected to move ahead in the innovation pipeline. The hunch is important because we have so much tacit knowledge that we cannot communicate directly. In the current state, its doubtful if our innovation AI could be developed to pick up on this. 

Culture is extremely hard to put a finger on, let alone manufacture. Because culture is hardly ever logical. Again, there is a certain human touch. Different cultures breed different ideas. Toyota’s vision of the perfect SUV is not the same as BMW’s. In the grand scheme of things, BMW might make choices that are not universally optimal when designing the next SUV. But they are optimal to BMW’s culture. We might even say that they are an expression of the BMW culture. Innovation needs the human touch.

Could your innovation department benefit from AI? Where would you implement it? If this is something you’d like to share your view on, or even better, experiences with, I hope you’ll reply to this email and get in touch.

I’ll sign off with a quote from Nick Bostrom. He is the author of “Super Intelligence“. In the book he predicts that AI will surpass human intelligence taking it to a level, that we as humans will not have the means to understand. Which in itself is an extremely scary thought if you ask me, but also interesting from an innovation perspective.


Think about it: Machine intelligence is the last invention that humanity will

ever need to make. Machines will then be better at inventing than we are.


Will this be the final innovation we ever need to make?

Leave a Reply