friendly-robot
Technology Planning and Analysis

AI: What's there to fear when it's already here?

This is a contributed piece by Jarred McGinnis, UK managing consultant at Ontotext

In the last decades of the 19th century there was a ‘War of Currents’ between proponents of AC and DC electricity transmission. There was widespread public apprehension of this new technology, culminating in the now infamous electrocution of a circus elephant at Coney Island to demonstrate quite how dangerous electricity was for domestic use. The absurdity of the situation was obvious to subsequent generations, and unimaginable today.

Today we’re fearful of different technology, yet the situation strikes me as similar. Led by popular figures like Elon Musk and Stephen Hawking, the public are convinced that the robots are out to kill us and we must spend millions to stop this deadly research gaining momentum. There is real concern among experts in the fields of machine learning, natural language processing and semantic technology, component technologies of ‘AI’: is the public being turned against research already improving our digital lives? Fear-mongering news headlines ignore AI technologies’ exciting capabilities for bettering our society through improving financial fraud investigations, advancing pharmaceutical research, or exposing political corruption.

What do we mean by AI?

AI is an umbrella term; for the purposes of this article we’ll focus on semantic technology, linked data, context-aware computing and how it relates to content management.

We’re all aware of the exponential growth of content online and the challenges that entails, and this is where semantic technology comes in: by marrying text analysis and content tagging with graph databases, computers are able to become ‘smarter’ than ever, understanding not just what text says, but what it actually means.

An example will help us here. Read the sentence “Mark Carney is governor of the Bank of England” and then answer the question “Where does Mark Carney work?” Amusingly, for a computer this is a difficult question to answer because the geographic location of Mr Carney’s office is not explicitly stated. To humans this seems nonsensical. Semantic technology improves the situation, enabling the computer to be ‘smarter’ by giving it the ability to infer. With semantics, the computer is able to understand contextual information on this “Bank of England” entity to infer that Mark Carney, in all likelihood, works in London.

This is not conjecture, this is already here: companies like Google and Facebook have been using this technology for years. The BBC and the Financial Times use Ontotext to enhance their news services by providing richer, deeper content suggestions. This is a real-life example of computers being trained to ‘think’ with less human input.

What does the future of the web really look like?

The core concept behind all this is linked data. As Tim Berners-Lee describes, the idea is for a semantic web where links go far beyond the document-to-document links we have today. We wouldn’t have to search for a fact by searching for a document, then searching for the fact within that document - we’d be able to go straight to what we want.

Digital content is organised into files and folders, not because that’s the most useful way, but because of a technological hangover from an archaic analogue age: grey filing cabinets, ring binders, paper labels. It’s absurd if you think about it.

Semantic operating systems can change this: they’ll be context-aware, understanding what your content is about in a more nuanced way. Rather than playing ‘keyword bingo’ to find those holiday snaps, your computer will be able to leverage existing metadata to organise them for you, and serve what you want to see more accurately.

I’m forecasting a change in how we interact with digital content thanks to semantic technology. We could revolutionise investigative journalism, inferring relationships between corporate employees and political lobbyists based on public data. We could improve our educational system by processing student work through semantic tools to assess more accurately and iteratively improve the syllabus accordingly.

This is all within grasp. The media frenzy about “AI killer robots” is frustrating, a hypothetical distraction from the tangible social benefits that context-aware computing can bring. These exciting applications fall well outside the specialisms of popular scientists who are capturing the national imagination with their doom-laden warnings, and that needs to change.

PREVIOUS ARTICLE

« Coupa, Nutanix, Okta and Twilio test 2016 IPO waters

NEXT ARTICLE

This month in tech history: January - TIME's Machine of the Year »
author_image
IDG Connect

IDG Connect tackles the tech stories that matter to you

  • Mail

Poll

Do you think your smartphone is making you a workaholic?