Digital Sucks. But, right now, AIs are helping us making it better

In films, AIs are homicidal machines come to kill us. In reality, they're tireless assistants solving the problems we don't have time for…

On the surface, AI — Artificial Intelligence — seems like the poster child of Digital Sucks. Culture has been warning of us of the danger of AIs since before I was born. Captain Kirk jousts with homicidal supercomputers, and Skynet becomes sentiment and sends armies of Terminators after us. The machines revolt, and use us as batteries.

They are sentiment. They rebel. And they have a plan:

But, you know what? That’s the purpose of science fiction, to explore what emerging technologies could be, and that focuses our minds on what they should be — and then we get a better result.

Digital sucks. Let’s make it better.

AI doesn’t look much like those sci-fi fantasies right now. Sure, we have talking assistants in our phones, but AI is already at work in other ways, at least in the more simple form of Machine Learning. And the phone in your pocket is already learning.

Teaching machines to learn

Open your Apple or Google Photos app and you’ll find movies of great memories from your life, made without human intervention. Microsoft’s Cognitive Services don’t just use face recognition for photos, but they can also make relatively accurate assumptions concerning age and mood of the respective person in the picture.

Machine learning allows computers to recognise the content of our photos, spots what’s good, or the people that matter, and pull together something that touches us emotionally. Apple has been selling us on this AI without going anywhere near the phrase:

What machine learning is allowing is AI systems that assist and complement human ability. Photography is still a human matter – we chose what matters to us, and take photographs of it. The AI kicks in when we generate such volumes of photos, as many of us do now, that we can’t organise them in a meaningful way. The labour of tagging, sorting and collecting can be assigned to a fast, tireless AI, and we can just enjoy the result.

AIs against terrorism

A more dramatic example is the way two big tech companies are now deploying AI to look for terrorist content on their networks. Here’s Facebook’s approach:

We want to find terrorist content immediately, before people in our community have seen it. Already, the majority of accounts we remove for terrorism we find ourselves. But we know we can do better at using technology — and specifically artificial intelligence — to stop the spread of terrorist content on Facebook. Although our use of AI against terrorism is fairly recent, it’s already changing the ways we keep potential terrorist propaganda and accounts off Facebook.

They’re using AI for:

  • Image matching
  • Language identification – spotting trends in the language extremists use
  • Removing terrorist clusters – that can be identified via AI
  • Preventing Recidivism

In these cases, the AI is the identification agent, bring it to human beings for review and action. Facebook is not alone in this approach; Google is doing much the same:

First, we are increasing our use of technology to help identify extremist and terrorism-related videos. This can be challenging: a video of a terrorist attack may be informative news reporting if broadcast by the BBC, or glorification of violence if uploaded in a different context by a different user. We have used video analysis models to find and assess more than 50 per cent of the terrorism-related content we have removed over the past six months. We will now devote more engineering resources to apply our most advanced machine learning research to train new “content classifiers” to help us more quickly identify and remove extremist and terrorism-related content.

Big data sucks. AI makes it better

Much of the value of AI will probably fall between the two extremes of personal support and international security. As the Harvard Business Review explored recently, machine learning has huge possibilities in product development:

Data can provide huge insights for companies, but making the most of the big data being generated is no longer possible without the help of machine learning. Artificial intelligence tools can help companies make better data decisions that improve the customer experience in real-time. And using data to drive more personalized customer experiences benefits customers and businesses alike.

More and more everyday objects, devices, machines or vehicles – in short, “things” – are interlaced through the help of chips, processors, cameras or sensors. Where computers have been the only interface for exchange of information on the internet, in the near future, nearly everything could generate, send or receive data:

  • Cars or machines collect data for maintenance and repair.
  • Robots in intelligent factories are detecting which production piece should be arranged at which point.
  • The refrigerator identifies what misses in the kitchen for roasting and puts this on the shopping list.

This is an example of how and why digital sucks – and then makes things better. Digital commerce and relationship tracking systems are great, because they generate data. Way back in 2011, we called our conference theme “Data Love”, because it was clear that digital tools generated huge swathes of data that we could use to make our work, our businesses and our lives better. But even then, we were starting to sound the warning siren of the next big problem: data deluges.

All too quickly, that volume of digitally-generated data becomes over-whelming – more than human agents can analyse at any reasonable cost or time-scale. Your customer database is huge, and disconnected from your repairs or services system, none of which can surface anything useful for marketing or product design, despite the fact that you know the data you need is in there somewhere.

How to solve that problem? Sure, you could throw people at it – but that’s slow and expensive.

The best option would be to bring together human beings and machines. Both contribute their very own talents: Humans are creative and innovative, emotional and empathetic. Computers are unsurpassed regarding their accomplishment to recognise patterns in massive amounts of data in an incredibly short space of time. Now it’s all about to reinforce human beings and machines concerning exactly these talents as well as to combine them specifically.

If we learn to cooperate with machines in a good way, we’ll be invincible. Together we’ll have the chance to finally solve huge problems of humanity. Artificial intelligence will help us controlling environmental risks much better, using resources more wisely or also reducing the number of traffic fatalities radically. But it’ll also improve the lives of individual persons significantly.

With self-learning AI agents, who sift and surface the most interesting elements of that data tirelessly — and with ever-improving skill as their machine learning algorithms gain more insight as they’re exposed to more data.

Sure, this sounds utopian. You might even question how it fits into our theme that “digital sucks”. Well, that’s because we’re on an upswing of the cycle. New technologies beget new opportunities which beget new problems – which beget new technologies to deal with them. It’s a cycle. And digital only sucks when you’re in the downswing of it.

Make the present better, and let the future take care of itself

As our speaker Bruce Sterling pointed out recently, science fiction isn’t always great at spotting the real dangers of a technology:

I have to protest that you will never actually get a “better future” from writing better about “the future.” No, that doesn’t work. That’s not how the passage of time actually affects a civilization. It’s like expecting a better past because you write a better history.

Sci-fi is great for improving the present, because it allows you to mindfully create with cutting edge technology, with an eye to its potential downsides. The unexpected downsides of the technology we’re using now are, by definition, unexpected. They are the unknown unknowns. We can deal with them in the future. There’s no doubt that at some point in the future some element of AI will indeed suck. But right now, at this point in history, we can enjoy the benefits it’s bringing us in sorting out the digital messes of the past.

There are plenty of great examples out there already. Heck, AI is even being used to make the cesspit of newspaper comments better. There’s a boon to public discourse…

AI is here. Apple is building machine learning into its machines. Google is realigning its business around it. Microsoft works on democratizing AI – so, all people are getting access to the intelligent tools and applications they need for the enhancement of their everyday life. Can you find some great new uses for it?