Convergence or Collision? A.I. and Content Marketing, Part 1

AI in Content Marketing
AI in Content Marketing

Historical terms are hard to ascribe to a topic like artificial intelligence. When you try to make comparisons, say to Henry Ford’s Model T assembly line and how it helped create an entire universe of mass production, the concept of A.I. – even in these early stages – dwarfs anything about a River Rouge, once you weigh the implications.

Has there ever been a more profound technological evolution than the one we’ve seen in the past 30 years? But we seem to be at a momentous moment, where the very phrase artificial intelligence lands on the ears like a theat. It’s intimidating to think that abstraction once just chewed over in philosophical bull sessions in your dorm room are now very, very real, and about to possibly – no, undoubtedly – impact your job as a marketer. And impact your very, very real paycheck.

Is this the dawning of a new era, full of promise and broadened horizons? Or a death knell for the human imprint? Is this the new Big Bang punctuating our existence? Birth or death? Convergence or collision?

We might have thought that, as content creators and strategists, our trade would be pretty safe from too much change. That ship has sailed.

The machines are already cooking up content

“While automation will eliminate very few occupations entirely in the next decade, it will affect portions of almost all jobs to a greater or lesser degree, depending on the type of work they entail,” according to McKinsey Quarterly.” Or striking closer to home, Gartner said, “By 2018, 20% of all business content will be authored by AI.”

The content that’s being drafted by AIs isn’t necessarily anything human beings will miss. But it shows us how writing and authorship aren’t numinously insulated from those soulless machine learning systems.

Interestingly, the harder the intellectual process — like calculus, financial market strategy, and language translation — the easier it is for a computer to execute, while the instinctive and reflexive aspects of the human brain — vision, movement, perception — are incredibly difficult for machines to emulate. Or as computer scientist Donald Knuth puts it, “AI has by now succeeded in doing essentially everything that requires ‘thinking’ but has failed to do most of what people and animals do ‘without thinking.’”

What’s all this mean for content marketing? Will AI mostly replace manual content creation? It sounds ridiculous to be referring to it as “manual” content creation, doesn’t it? By manual, we mean tapped out by an actual human being, though it almost makes it sound like the scribbling we’ve been doing all these years was akin to making handmade furniture or craft beer.

What’s next? “Artisanal content creation”?

Quick, let me throw a trademark on that: Artisanal Content Creation™.We may chuckle at it right now, but who’s to say what we’ll think in 2025?

ANI, AGI, ASI, oh my!

First, let’s wrap our heads around the three levels of AI that are either here, or being projected in the still-foggy future: Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI), and Artificial Superintelligence (ASI).

Sometimes called Weak AI, Artificial Narrow Intelligence (ANI) is AI that specializes in one area. Like an AI that can whip the human world chess champion, but it’s a one-trick pony. This is also the kind of AI you find in cars navigational systems, smart phones, email spam filters, Amazon and Netflix recommendations, Google Translate, Google Search, and your Facebook newsfeed.

Then there’s Artificial General Intelligence (AGI), AKA Strong AI, or Human-Level AI. AGI refers to a computer that’s as smart as a human at everything — a machine that can perform any intellectual task we can. Creating AGI is far harder than creating ANI, and we’re yet to get to that threshold. When hundreds of scientists were asked when we’d likely achieve an AGI, their median answer was 2040.

Finally, there’s Artificial Superintelligence (ASI). Oxford philosopher and leading AI pundit Nick Bostrom defines ASI as “an intellect that is much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills.” The range encompassed by ASI could include computers just a little smarter than you or me, to machine intelligences that are trillions of times smarter.

So how smart is (super)smart?

How big is the gulf between AGI and ASI? Here’s where the sci-fi scenarios really kick in. Thanks to recursive self-improvement, an AI that reaches a certain plateau – say, that of a village idiot or average politician – will just keep on following its programming, which drives it to improve in own intelligence. Since it’s already faster at processing routines and teaching itself than a human, the next set of leaps may happen in not years, or months, or weeks.

An AGI may rocket upward in intelligence in only hours, reaching the beyond-our-ken levels of an ASI system. The term attached to this, Intelligence Explosion, sounds appropriately ominous.

Some AI prognosticators say you may step out for lunch, leaving behind your AI assistant, who has the brains of a Bender. When you get back to the office, you suddenly find you’re taking orders from Ultron.

And there’s no escaping a mention of the Singularity, of course. Which has lent its name to at least one godawful movie which asked the compelling question, “How low will John Cusack stoop for a paycheck?”

The actual question that gets posed vis-à-vis the Singularity is this: Will human beings and machine intelligences merge into a conjoined entity?

Ray Kurzweil, who invented the term, defines it as a specific moment, “the time when the Law of Accelerating Returns has reached such an extreme pace that technological progress is happening at a seemingly infinite pace, and after which we’ll be living in a whole new world.”

Will a content creator still have a job in that world? It’s hard to say. What’s more concrete is the fact that artificial intelligence is already stepping up its content game – and we should be ready.

A change in the Weather

A few years ago, IBM bought the Weather Company’s assets. Sounds pretty random, right?

Yet they already had an alliance in place – one that allowed IBM to sell the Weather Company’s data services and incorporate that data into Watson’s services. The main idea behind Watson is that users can ask it questions using everyday language, such as “What’s the weather like in New York tomorrow” and the system will provide an answer.

This newer acquisition gave IBM control of a Weather Company mobile app too – an app that’s installed on millions of smartphones, all over the globe. No small feat.

This illustrates just how quickly an AI can become ubiquitious, creating new interactions between content and consumer. And if highly personalized, on-instant-demand AI works for the weather, there’s no reason it can’t work for content personalization.

Automating Authorship

Buckle up for the first big truth: AI is already being used to create content like Forbes earning reports generated by NLG’s Quill. These reports are very data-driven, and fit neatly into the parameters of an AI (if the data and rules are correct, the prose will be produced) – but they don’t have any actual deep-dive analytical bent. That makes them perfect for an AI to churn out, as there’s no subjective judgment involved, only hard fact recitation.

The Associated Press creates its quarterly earnings reports via AI. Turning out a quarterly can be demanding, stressful, and monotonous as hell. The accuracy and speed an AI can bring to the job makes sense. So AP partnered with Automated Insights to auto-scribe its quarterlies using the vendor’s Wordsmith platform.

But wait! AP also uses it to generate articles, news reports, and more. According to Automated Insights, their system can churn out 2,000 articles per second, if it had to. Before implementation, AP was creating quarterlies for 300 companies. Now, it’s doing 3,000 per quarter. Some will add a human grace note or two, with edits to the computer-generated material or by doing a separate follow-up item.

Both AP and Automated Insights claim nobody has gotten a pink slip as a result of AI, while fewer errors are being logged versus human-generated copy.

Sure, it sounds like a win-win. For now.

Not all words are created equal

The ugly truth, presented in this study on digital journalism, is that most readers cannot differentiate between content generated by software and content written by journalists.
We aren’t saying that there aren’t slight differences – readers find machine-written content to be more boring and descriptive, but they also believe it to be objective.

So human-written content may be less boring and “significantly more pleasant to read” – but to the average reader, the author could be human or machine.

‘Smart’ tools permit perfect timing

Better late than never? That’s not an excuse that stands up for anybody in modern digital marketing.

Let’s face it, punctuality just isn’t everyone’s cup of tea. But the consequences of not having great time management skills may be a thing of the past, at least as far as social media is concerned. As automation seeps into scheduling, social media management tools are in every marketing team’s back pocket. Hootsuite, Buffer, Sprout Social – these tools allow you to write (or machine-generate) posts for your company’s various social media accounts in one tidy dashboard, then schedule them at intervals of your choosing for the indefinite future. Automation at its finest!

Right next to their favorite scheduling tool is every marketer’s preferred automated way to track data. Conversions, lead flow, web traffic – these are all bread and butter to today’s digital marketer. From incoming web traffic to individual topic performance, everything is tracked in terms of links, clicks, shares, and engagement.

The data is complex and often too vast in quantity to deal with manually, but without this data, any marketer would be lost – and pretty soon unemployed. Automation is making these tasks simple for we harried marketers, and not a minute too soon.

Notifications have been automated; marketers can be notified if their brand is mentioned in the news, for instance, or if an article hits a certain popularity threshold. This goes far beyond vanity, as certain mentions can prove the value of a new marketing campaign. Being mentioned on social media might give a brand the right traction to launch a new campaign, or show them how to optimize it. Being alerted right away can have big impact when mere moments can make a difference, allowing marketers to capitalize on what’s hot or trending and tweak their campaign accordingly.

Automated ways of updating editorial calendars can cut down on hours of wasted time, leading to better organization and reducing human error. Even correspondences between team members have become automated, with dashboards that allow higher levels of transparency and fewer mistakes.

Even at their most sophisticated, these sophisticated forms of content automation are not AI tools – yet.

The great debate: Is it AI or automation?

We’ve discussed both artificial intelligence and automation – so how are these two categories defined? The lines are blurry and opinions vary, but we can take a closer look at the debate and the questions that guide it.

Automation is defined as making hardware or software capable of doing things without human intervention. Artificial intelligence, on the other hand, is the making of intelligence machines that can mimic and eventually supersede human behavior and intelligence.

“It is the characteristics that the final tool exhibits that are relevant.” – Hackernoon

When taking this approach, it’s important to dissect how proactive or reactive the tool is that you’re using. Many tools today are reactive, which means they respond only to external stimuli in their environment. This means their reactions are not deliberative or intelligent, which means that they aren’t really AI. A proactive tool which shows deliberation, on the other hand, takes the initiative without requiring any external stimuli.

What are a few key questions to ask in determining if a tool is AI?

Can the tool learn and adjust future plans for better success?

If a tool is truly “artificial intelligence,” it should be capable of bettering itself. The more it performs an action, the smarter it becomes. This differentiates AI from automation.

How autonomous is it?

When you look at self-driving cars, there are 5 levels of autonomy. These vehicles use a combination of sensors, cameras and AI to become ‘driverless’. As of 2019, we’re only at level 3.

As one observer put it, “Self-driving cars cannot decide that instead of taking us to point B they would prefer to take us to point C. When that starts happening that is when they are truly autonomous.”

How creative is it?

The idea of an AI being truly creative may be anathema to some, but a lot of human “creativity” is just clever imitation of others’ originality and genius. You only have to check out the pop charts for proof of that. As IBM points out, AI has already “helped write pop ballads, mimicked the styles of great painters and informed creative decisions in filmmaking.”

“How creative” an AI is depends on the perspective you, as a potential user, is bringing to the question. Creativity sometimes shows itself in blistering excess, but at other times in winking, almost imperceptible subtlety. Let’s just say that to truly qualify as AI, a platform or tool should show some ability to create something that wasn’t there before, whether it’s a predictive insight or a fresh angle on content generation.

Or its own language. Which was “creative” enough to even scare its developers.

How connected is it to other systems?

Can it plug into different platforms and systems to gather the data it needs, and even converse with other AIs to put their digital noggins together to plot global domination refine and improve outputs and outcomes? So much of “intelligence” in any form is the result of being able to synthesize data and other inputs, and a true AI has to be designed to connect and gather what it needs to accomplish that job.

Evolve or perish, publisher?

AI will, however, continue to be more cost-effective, efficient, and faster than humans could ever be. So for better or worse, AI will increasingly be used as a platform for creating content, endangering human jobs in the process.

“There certainly will be job disruption. Because what’s going to happen is robots will be able to do everything better than us. … I mean all of us…I am not sure exactly what to do about this. This is really the scariest problem to me, I will tell you.”Elon Musk

But maybe the solution is, as some suggest, for human content creators to move on from the kinds of tasks that AIs can execute. Or integrate AI imaginatively into their work so they can evolve and elevate their content creation game to places we can’t even imagine yet.

After all, Herr Gutenberg may have put more than a few illuminated manuscript makers out of work, but nobody today can argue with the upside. But nobody alive in 1439 could predict how far a disruptive new tool like the printing press would take us, and only saw its dangers to the status quo. Right now, we might be dreading AI the very same way: without a real idea of its eventual benefits.

Spread the word: