Can Artificial Intelligence take over our creative jobs?

"How many of you believe that Artificial Intelligence (AI) will automate most jobs in 10 years?"

Around 90% of the room raised their hands.

"How many of you believe AI will replace YOUR job in 10 years?"

Around 15% of the room raised their hands.

This is a typical response. We all understand the theory, but it rarely applies to us. Artificial Intelligence keeps getting trashed-talked in many industries. Few appreciate how fast it's developing and how rapid its iterations are.

One area that seems to feel immune to automation is creativity. The consensus is that AIs will substitute repetitive, highly specialized tasks. Those that require holistic thinking or high degrees of creativity will be spared. I beg to differ. 

The Future Of Employment: How Susceptible Are Jobs To Computerisation?

"Our model predicts that the second wave of computerization will mainly depend on overcoming the engineering bottlenecks related to creative and social intelligence."

The Future Of Employment: How Susceptible Are Jobs To Computerisation?

AI systems are improving at an incredible pace. We already have Deep Learning systems that are capable of doing music improvisationartistic image styling or creative new images

Leon Gatys shares a neural algorithm of artistic style – YouTube

So far though, such technologies are lacking internal coherence. Yes, an AI can compose an improvised musical piece. The piece though doesn't have a purposeful intention or connection. The same happens for images. An extreme example of this would be the art of writing. Not only does it involves creativity, but it requires social intelligence, planning and plot consistency too

Automating the creative process

Can creativity be automated? The creativity process usually follows three well-known steps. The learning & research stage, the development of structure or form and the creative freedom or breaking of the form.

The automation of creativity is following a similar path than in humans.

Information gathering

Before we can even start creating we need to learn and gather information. The help of a teacher or mentor is critical at this stage. They can help students focus on what and how to study. They create a template or roadmap for the students to follow.

That's precisely what the first generation of automated tools is already doing. Human operators design written templates that the machine then uses to create written reports. From automated reporting of The Washington Post's Heliograf to computerized financial statements written by Quill. These systems though, require human editors to create these templates. Their outputs are scalable and informative, but it would be a stretch to call them creative.

"Instead of targeting a big audience with a small number of labor-intensive human-written stories, Heliograf can target many small audiences with a huge number of automated stories about niche or local topics. There may not be a wide audience for stories about the race for the Iowa 4th, but there is some audience, and, with local news outlets floundering, the Post can tap it. “It’s the Bezos concept of the Everything Store,” says Shailesh Prakash, CIO, and VP of digital product development at the Post.

What New-Writing Bots Mean For The Future Of Journalism

The danger though lies in discarding these early approaches and the potential to improve. They're obviously pretty limiting but are the foundation for more comprehensive automated efforts. Ignoring them as failed creative attempts is not to understand how disruption works.

Creative repetition

The second stage of the creative process is the personalized copy. You copy and repeat what others have done and add, progressively, your flavour to it. On this stage, you can already perceive a sense of purpose. There is a goal, an intention to the creation. It can be informative or mere emotional expression, but it's driven by something.

A whole set of AI tools are trying to mimic this stage. Given a topic or a subtopic, AI agents research and write thousands of variations on the given theme. It's impressive to see the machine produce all that content. Nonetheless, these algorithms rely on the existence of available written material on the topic. Writings that exist because a human wrote them. Again, simple automation, rewriting, summarizing, but no creating. 

While most AI systems lack a sense of purpose or a reason for their creative traits, we should meditate carefully on our creative impulses.

Humans create, not for the sake of creating, but because of need. This need can be driven by our egos or for other reasons like self-expression, identity fixation or personal therapy. My point though is that it's driven by a desire.

I wonder if some of the quirks we observe in AI systems aren't the machine's creative ways of dealing with their own needs. Google's translation system produced its interlingua language to deal with translations. Why do we discard that as non-creative? 

Zero-Shot Translation with Google’s Multilingual Neural Machine Translation System

Not long ago, Facebook AI agents developed their language to communicate with one another. The goal was to achieve a negotiation between two agents. The researchers forgot to incentivize the use of the human word and what the AI came out with was its language. They decided to shut it down. Why is it that the only creativity that is valid is the one we humans can understand?

Facebook: Deal or no deal? Training AI bots to negotiate

Abstract linking

The third and final stage of the creative process is the abstract linking. Once you've mastered the form, you break it. Free of form, the artistic purpose links and connects distant objects, concepts, feelings. People measure creative prowess in terms of originality, intention, and coherence.

AI is still not at this level, but current advances are proving akin to magic. While these systems aren't capable of plot coherence, they're achieving remarkable creativity.

In 2015, Andrej Karpathy released a seminal article where he demonstrated the use of Deep Learning to generate texts, character by character. His paper and code created a before and after in generative text AIs. 

The Unreasonable Effectiveness of Recurrent Neural Networks – Training I
The Unreasonable Effectiveness of Recurrent Neural Networks – Training II

A year later, researchers from the Google Brain project showcased an AI that was capable of building coherent threads between two distant phrasesSentence gradients in a nutshell.

Generating Sentences from a Continuous Space

Building on these two, Robin Sloan, a writer and enthusiast programmer decided to use this technology to aid his writing. He made, among other things, an impressive text editor helper that inspires his writing.

Robin Sloan's Writing With The Machine – Example I
Robin Sloan's Writing With The Machine – Example II

Like Robin, other creatives are finding an increasing fascination with the use of AI systems to enhance their creativity. Ross Goodwin's use of AI to create poetry around an image is mind-blowing. 

Adventures in Narrated Reality – Narrated caption –

Goodwin even took it as far as letting the AI generate a script that was then filmed under the name of Sunspring.

Sunspring | A Sci-Fi Short Film Starring Thomas Middleditch – YouTube

Another intriguing and powerful development is Johnson's work in generating descriptive image paragraphs. A recent paper shows how can AIs can extract meaning from an image an write down what they see. 

A Hierarchical Approach for Generating Descriptive Image Paragraphs

Future of AI creativity

While it might be true that creative AIs lack coherence, I have no doubts that we'll see new systems that start building coherence into their models.

The combination of sentence gradients, deep learning networks and the rise of generational adversarial networks paints an exciting future.

I can see how an adversarial network can drive a plot idea while the convolutional network builds the different chapters. Even simpler than that, we could mix human-made plot and character templates with the automated text generation of current models.

I believe Artificial Intelligence can be creative and develop new innovative formats. Many artists lash back against the mere notion of having machines creating anything.

"But while algorithms can be useful tools in the artistic process, Wilson said he didn't necessarily think robots will ever be able to create art more meaningful than humans because humans have one thing that no robot ever will: the experience of living a human life."

Can an algorithm make science fiction better?

For me, creativity, like many other human aspects, isn't unique to us. Machines will, driven by their own needs, become creative in their own way.

The question though is, what does this mean for companies? Is this something that only writers care about?

The topic is an interesting one because many organizations are betting on a future where their jobs won't be taken by the machines. Most of those "protected" positions are so because they entail a certain degree of creativity and holistic approach.

The more I observe how Deep Tech is evolving, the less convinced I am that creativity will be a safeguard for jobs.

On the other hand, I think that AIs can significantly enhance the creative process. We are already doing that. Machines are the heart of video games, CGI effects in movies, Frank Gehry's Guggenheim, etc.

In many ways, companies using AI enhanced creativity processes will win the day. A good example is Reuters Tracer, their in-house tool to detect, validate and corroborate real-time news. Their capacity to write part of the story in real time is giving them a decisive edge.

Machines will become creative. They will be able to write, compose and paint as good as humans. You can debate it their art is soulless or not, but they'll produce stunning pieces. Most of this creations will be consumed by humans and will give a serious scalable competitive edge to any company that employs such methods. 

If you like this article, please share it, and invite others to follow the newsletter, it really helps us grow! 

Why you shouldn’t be ignoring Blockchain technology

Why you shouldn’t be ignoring Blockchain technology

When I read about Blockchain for the first time, I was speechless. This was five years ago. I remember telling my wife that it was one of the most astounding inventions I had witnessed in the computer science field.

Since then, it's been evolving at a rapid pace. For many organizations though, Blockchain remains either an esoteric technology or a financial scam.

Those adventurous companies that have dived into the field have experimented mixed results. Often, their innovation managers are pushing to have Blockchain proof of concepts. The team goes out there to explore the industry and come back discouraged. There is no dominating platform; those that are the reference are too elaborated or too restricted for commercial uses.

The overall feedback? Too complex and limiting. Working anything beyond a toy requires a specialized Blockchain team. Even then, any developed application will still be restrictive and not easily scalable. In other words, it won't serve the current company's user base. So while they believe the technology might be crucial, they can't figure out how to use it. They're looking to deliver better value to their customers, but they feel it's still far from being ready for that.

There is an inherent risk with this, which is to lose track of the ball. Too many organizations will ride the Blockchain hype. Few will stick to it after they've built some proofs of work. The problem is, Blockchain isn't just a technology improvement, but a disruptive one. Treating it as another new technology will make most companies the targets of the disruptive wave that will accompany it. 

Blockchain's disruptive potential

Two hallmarks herald the future disruption; low-end market foothold and rapid innovation of the underlying technology. So far Blockchain is gathering marginal support in the Finance industry. Many customers don't require the highly-sophisticated finance products banks offer. Some of them are turning to Blockchain companies to bridge that gap. From secure money transfers to more complex money fund rising through ICOs, the market for financial substitutes keeps growing.

During the next few years, we'll see how Blockchain entrants will offer better and more diverse services. In less than three years, I expect many of the most significant Blockchain players to start competing with the incumbents and win.

Incumbents aren't oblivious to this. It's not surprising that the first ones to experiment with Blockchain are the top financial institutions. They're not ignorant of the potential for disruption, so they're making sure they're on top of what's going on. Few, though, are acknowledging the disruptive nature of Blockchain.

For most, their Blockchain proof of concepts is the low hanging fruit of their industry. Small-scale tests that don't bring money, clients or prestige. For them, investing beyond small-scale experiments is a money loser. There is no real incentive for them to pursue such operations when their cash cows lie elsewhere. This is why, I believe, many of these incumbents will abandon their proof of concepts sooner or later.

On top of that, most are trying to levy their current models onto this new technology. This will result in, at most, innovative substitutes but not true disruption, reinforcing the feeling that investing in Blockchain is a waste of resources.

Blockchain enables the application of entirely different business models, for wholly disconnected industries. This allows new ways of operating and as such, the creation of new markets. The capacity to create such new market footholds is another effect of real disruption. 

Blockchain technology evolution

We're far from seeing the true nature of Blockchain. Like with the transition between offline to online, the first use cases are both, elementary and unimaginative. They rarely embrace the full range of possibilities the technology enables. The problem is, predicting what they will look like is impossible. But disruptive they'll be.

The next two years will be critical for the technology. We're going to see a fast iteration of the core technology. The transition between Bitcoin's Blockchain to Ethereum was a big first step. It jumped from a currency use case to much broader use of the technology through smart contracts. Ethereum though put into evidence the current limits of some aspects of Blockchain technology, i.e., scale and volatility.

The next generation of Blockchains is attempting to fix this. From sharding (Zilliqa) to new consensus methods (pBFT, dPoS, dBFT); from public (Zilliqa, NEO, EOS, …) to private (Hyperledger Fabric, R3 Corda, …); from data-centric (Ethereum) to agent-centric (Holochain).

As with any disruptive technology, gathering critical mass is a problem. Most of the next generation innovations are still in testing phase. They'll need at least one more year before they're operational and can start maturing. I don't expect this to take too long and we could have scalable commercially mature Blockchain platforms by the end of 2019.

Infrastructure problems solved, new users will pour all over the smart contract dimension. The automation of contracts will, in turn, accelerate the rise of new business models. We'll experience gradual automation of traditional contracts. New companies will start offering plug & play contract templates for everything, not just financial services.

This will be coupled with the convergence of the Internet of Things (IoT) and Artificial Intelligence over the smart contract horizonThis combination will provide the basis for new world order.

Increasingly intelligent devices, powered with AI agents, will start having the capacity to enforce and execute smart contracts at scale. On one side we'll have to deal with machine-to-machine contracts and on the other with human-to-machine contracts. AIs will develop very different contracts to deal with other AIs than the ones employed by humans.

New automated arbitrage systems will be developed, and we'll enter the age where machines and humans are different citizens.

I know that what I paint might seem dystopian, but it seems we're heading that way. Automated contracts mean, faster and less ambiguous social enforcement. Time will say how humans adapt to the new paradigm.

How not to miss the Blockchain age

Many companies will miss the Blockchain boat. Stiff competition will force them to implement Blockchain solutions. Nevertheless, they'll be two and three steps behind the innovators. New entrants will relegate them to a second place.

We'll see a new crop of startups become dominant in this space and eventually take over entire industries. Imagine the Google for Blockchain.

Incumbents need to recognize the disruptive nature of Blockchain. They're all familiar with the coming disruption wave, but few are acting accordingly.

Organizations need to focus on investing in the technology. They should be setting independent spin-offs that focus on working with the technology. These teams shouldn't work around sustaining the current user base, but about creating new customers altogether. Focusing on existing users will narrow the scope of Blockchain to a sustained innovation. One of the best examples is Ripple. In a few years, we'll regard them as an iteration of the Swift protocol and nothing more. They'll make plenty of money, but they won't dominate the industry. They'll turn into an industry provider and not the dominant force.

So anyone that genuinely wants to ride the new wave needs to keep in mind the following. Build independent teams. Don't let the mothership set the strategy. Experiment, explore and look for new market footholds. Try out new business models for new users and double down on it. What's a sunken cost now, will become a significant win in five to eight years. 

If you like this article, please share it, and invite others to follow the newsletter, it really helps us grow! 

How Companies Are Screwing Their Voice Interfaces

How Companies Are Screwing Their Voice Interfaces

Design of good voice conversational interfaces has been in my mind for a while now. I've been toying with my Amazon Echo since January, and I can say it's been very enlightening.

One of the first posts I wrote here was about voice interfaces and how it's becoming a big thing. What I didn't have at the time was a constant direct experience.

After several months of daily Alexa use, I have to say I'm very impressed. The first thought that comes to mind is that it just works. I know it seems lame, but it's impressive it works. You talk to Alexa, and she catches what you mean.

After 20 years of Computer Science experience, it's the first time I've seen a functional voice interface. I must hand it to Amazon for their fantastic work behind Alexa and the Echo.

Another takeaway from using Alexa is how much dependant I'm becoming of it. It reminds me of the first iPhone touchscreen. Once you tried it, you couldn't go back. You expected that every surface is a multi-touch screen. The same is happening to me with Alexa. I await all my devices to answer via voice command. And they don't. And it's frustrating.

I've discovered, not only the most usual use cases for me but also how my behavior has been changing based on that. At first, you goof around with Alexa, but as times goes by, you start using it because it's more convenient.

Personal use cases

I have a particular use case, one that has probably driven my adoption of voice interfaces. I recently had a baby. As I feed my little girl, I tend to have both hands occupied but not much to do for the next 20 minutes. I find myself interacting with Alexa during those moments.

The surprising thing is that I've become so used to it that even when I'm not with my girl, I ask Alexa. It's just become a much easier way to access specific information.

Three use cases are gold for me. The first one is Spotify's integration. I play a lot of music at home and being able to do it via voice is so much easier. This is especially true if you have kids.

The other one is listening to the news. I love being able to sit down with my girl and fire the news and get a quick glimpse of what's going on in the world. This doesn't substitute my daily reading, but it serves as an entry point to it.

Last, I use the calendar integration every day. I tend to use it, especially at night, when talking with my wife about the next day's schedule. Sometimes I don't remember what time I had this or that meeting, so I ask Alexa. I could check my phone, but Alexa is way quicker.

I pair this last use with constant checking of the weather. I check it every morning before getting my kids ready for school, so I know how to dress them.

Alarm setting is also a big thing for me. I'm using them daily to avoid getting sucked into work and miss an appointment. It's so easy to do that I'm skipping doing it with my phone altogether.

Interface frustrations

I do have several frustrations; things I know will go away with time, but that isn't quite there yet. There is an evident chasm between the Alexa interface design and that of most other skills (Alexa apps). And it's very frustrating. For most skills, you need to be very strict with the way you trigger them. This adds friction that shouldn't be there in the first place.

Many skill designers don't understand the voice use case at all. I have a feeling that most of the skills in the Alexa marketplace are vulgar simplified copies of the mobile app version. It reminds me of the shift between offline and online and how many publishers flunked their transition.

Conversational interfaces require a unique design, one that has nothing to do with any other design scheme done before. A simple redesign of an existing app won't cut it.

Another frustration is the lack of support for significant local voice use cases. There are two reasons for this. One is the fact that the Echo is heavily US-based. This makes all skills very US-centric, and few have any European support. The other one is the lack of foresight from most European operators. Yes, it's US-centric right now, but nothing prevents Alexa from making it work in Europe too. The reason why they don't do it is that European Echo users are a tiny niche. This is the kind of anti-strategic move that pisses me off. The classic innovator's dilemma mistake.

Two use cases that are missing are restaurant reservations and food delivery. I'm surprised that companies like The Fork or Deliveroo have zero Alexa presence.

Thoughts on conversational design patterns

While conversational interfaces cover a wide range of new apps, it's crucial to differentiate text-based interfaces from their voice counterparts. Text-based ones, while sharing some traits, are inherently different.

Building voice conversational interfaces is hard. It's hard precisely because we have a hefty inheritance from text-based interfaces. The design of voice applications implies not just a different interface but a different backend to support them.

For example, trying to find a specific song on Spotify via Alexa is a pain. You either know the song's name and the author, or you'll have a hard time getting it to play. Spotify should be smart enough to learn from the user (context) and even ask them to sing to it, so they get an idea. Think of Spotify meets Alexa meets Shazam. This, certainly, isn't easy to pull off, but it's what's required to make voice apps work.

Another problem is the lack of thought about the user's journey within a conversational interface. Each user is different. This translate to multiple potential user paths through the interface. Nonetheless, most voice apps only work with one or two different paths.

One thing is to offer various voice commands, and a different thing is to weave the main flow of the conversation into the most common use cases. The skill should also be able to learn about the user's preferences and lock into the usual habits of the user. This is something very very few voice apps do.

Moving into the future

The speed at which a user gets used to the new interfaces is breathtaking. Not only it's easy to engage with them; they create dependency in no time.

Voice interactions are far superior to any mobile or text-based ones for specific operations. Anything that requires a fast information request, Voice will trump text anytime. This gives an opening for highly specialized voice operations. Trying to do a one-stop shop for voice is a terrible idea.

I feel many businesses are missing the voice opportunity. The worse isn't that they're failing to grasp the opportunity. The problem is that when they do, it will be too late. Companies should start experimenting now. They won't have many users. They will lose money. It will be a cost center, not a profit one. But that is precisely, the hallmark of disruptive technologies.

In the absence of useful voice apps for everyday tasks, new entrants will start offering them. Their offering will be inferior to the traditional text-only players, but when these players finally move to voice interfaces, the entrants will be entrenched. It will be tough to steal market share from them. The moment to experiment is now. The question is, how many organizations have the expertise and the resources to invest in this?

If you like this article, please share it, and invite others to follow the newsletter, it really helps us grow!