Even now, in late 2018, the public’s perception of artificial intelligence is rooted in science fiction. A search on a popular UK tabloid newspaper for “artificial intelligence” demonstrates that around 50% of articles referencing AI also contain links or references to Robots. Even news articles that should be extolling the breakthroughs in deep learning algorithms – such as this one on Google AlphaZero in The Sun – will revert to calling it  “superhuman,” it seems the media can’t help itself in promoting a populist version of artificial intelligence.

Perhaps AI needs a rebrand. Even the name, artificial intelligence, alludes to fantastical creations – it’s straight out of the 1950s and 60s ideas and Lost in Space or Star Trek. TV and film have skewed our perception. Also the term has been hyped for years with little to show for it outside of Hollywood.

In reality, it’s only since 2010 the technology has snowballed. It’s a subset of a subset of AI that has got all the attention – Deep learning – and it’s evolving at a rapid rate. The most popular tool, TensorFlow, was only open sourced by Google in 2015. The results of all this collective effort are slowly encroaching on everyday life.

That’s the reason why the media is still stuck in the 1950s – it’s clear that deep learning is useful in many ways, but it’s not quite as sexy as cognitive robots.

Examples of deep learning include Google translate helping us overcome the language barrier, voice recognition in home assistants, product recommendations during online shopping, phones that unlock themselves through facial recognition. It’s all great, but there’s no poster child “killer app” for the technology.

It’s true value probably lies in “slightly better” decision making in all sorts of businesses – helping them to reduce waste or effort, improve pricing, upsell additional products – but “slightly better” doesn’t make a great headline. Generally what the media do is aggregate all the “slightly betters,” calculate the reduced effort, and do a back-of-a-postage-stamp calculation to figure out the number of jobs that equates to and hey presto, a perfect “Machines Will Take Our Jobs” headline.

How can we fix it?

Fixing it means understanding the technology.

Too much of the AI hype in the media is due to fundamental misunderstanding with the technology – how can a PR professional publicly promote your technology without understanding it?