CognitionX Community Drinks – AI in 2017 in Review

CognitionX summoned AI experts, C-level executives and Academics for a review of AI in 2017 and some punditry on what 2018 holds in store.

Between mouthfuls of mince pies here are the five top things our experts said:

The UK is full-steam ahead to be a world leader in AI. We have a good head start, a strong research and talent base and political conditions are favourable; we heard about the plans for a new government Office for AI to deliver on the Industrial Strategy. Discussion was lively about whether the UK could compete with Silicon Valley for funding and investment. Cambridge University’s Professor Jon Crowcroft pointed out that we may actually be avoiding many of Silicon Valley’s mistakes.

The AI world is intractable from macro social trends. Tabitha Goldstaub led the room in celebration that “2017 was the year it was no longer OK to be sexist and misogynistic”. She remarked how much positive pressure this puts on AI companies to diversify their development teams and product offerings, as well as minimise biases and their effect on underrepresented groups.
“2017 was the year it was no longer OK to be sexist and misogynistic”

2018 will about moving from Lab to Live. Karim Jalbout from Egon Zehnder was clear that companies across industry have seen enough successful proofs of concept. 2018 is now a race to master applied AI in practice, not just theory. That starts with the realisation that hiring a Chief Data Scientist is not enough when a company’s whole approach to data collection, storage and exploitation is sloppy.

Some key AI tensions might resolve themselves in 2018. For example, Mariana Pereira from DarkTrace raised the developer’s choice between AI processing on local devices (e.g. Amazon’s DeepLens camera unit) versus AI processing in the cloud (e.g. Alexa). Will one approach come to dominate? These are decisions driven by cost and efficiency as much as consumer preferences and data privacy regulation.

Explainable AI is just around the corner. Academia is closing in on some of the most awkward problems related to Deep Learning. Professor Crowcroft stressed how important getting this right will be to gaining public trust. The example of the de Havilland Comet jet plane crash provides a useful historical parallel. Until engineers could explain what went wrong, all UK production was halted. In that time Boeing pulled ahead in the US and still leads the market today.

From these discussions as well as the last 12 months, it is clear that AI is now crossing the chasm faster than any technology before it. We are now moving from experimental to mainstream. This is an exciting and important time to deploy in your organisation.

Published in

Leave a reply

Thank you! Your subscription has been confirmed. You'll hear from us soon.

Log in with your credentials


Forgot your details?

Create Account