Your browser is not supported. Please update it.

7 March 2019

Faux AI start-ups finally called out – fat chance of change

How apt that just days after MWC closed its doors to a discord of AI back-patting for a job well done, a simple survey stole all the headlines with the damning conclusion that not a morsel of AI was found in the products of 40% of so-called AI start-ups.

Let’s not kid ourselves, virtually everyone in this industry was already abundantly aware of such fallacies, so it’s nice to finally have some facts and figures to wave in the face of deniers. One of the most painstaking parts of our job here at Faultline Online Reporter is separating the phony from the disruptive, yet even by our high standards of skepticism, the 40% figure was a minor surprise, and we suspect once the storm dies down it will be business as usual for the faux AI crowd.

The findings come from a survey conducted by venture capital firm MMC Ventures of 2,830 purported AI start-up firms in 13 EU countries most active in AI.

Given the generally positive tone of the 150-page The State of AI Divergence 2019 report, it’s safe to say that MMC Ventures has inadvertently whipped up a storm. We have little doubt the wording intended to put a positive spin on finding genuine AI in 60% of start-ups surveyed, which perhaps reflects the naivety of venture capital firms. And to that end, it’s easy to criticize start-ups for lying about their products, when maybe we should instead point the finger at where the money is coming from – the VC world. By driving the frenzy to invest in all things AI, who can blame small companies for wanting to grab some funding?

Still, dressing up a product as AI, simply by injecting a new piece of programmable software and throwing in some generic algorithms, is something we have warned about time and time again. Often during interviews, we can probe executives to the extent they eventually let down their guard down by admitting some off-the-shelf AI algorithms are used, but always make sure to say these are used in conjunction with their own in-house developed AI algorithms.

“But what makes your product artificially intelligent?” is a question which more times than not conjures up blank faces.

“Over time, the distinction between ‘AI companies’ and other software providers will blur and then disappear, as AI becomes pervasive. Today, however, it is possible to highlight a sub-set of early stage software companies that have AI at the heart of their value proposition,” said an interesting snippet from the MMC report.

So, would a fair conclusion go something along the lines of the industry being on the verge of an AI market crash? Well, before we get accused of scaremongering, an in-depth report from our research arm Rethink TV thought as much last year, long before this week’s survey stole the show. It projects the global AI market to reach just $39 billion by the end of 2023 (significantly less than contrasting reports), giving rise to a bubble which is destined to burst – dragging with it millions of dollars of venture capital funding into the abyss. It concludes that most AI riches will become conglomerate exclusives, while only a handful of niche companies will thrive in specialist sectors.

While the focus has been on a tiny section of MMC’s report, there were also some insights into AI use cases, highlighting that technology and financial service companies are absorbing 60% of AI talent, which is in turn causing a “brain drain” in academic circles. Considering we would define AI itself as a technology, it seems strange to have technology as a separate sector entirely to, say, transport or healthcare, two of several verticals accounting for the remaining 40%.

More interestingly, it finds that professionals in the technology, media and telecom sectors spend 36% of their time collecting or synthesizing data, while finance and insurance tops the list at 50%. As such, in approximately 60% of occupations, at least 30% of constituent activities are technically automatable, says the report citing McKinsey Global Institute data, although automation is another kettle of fish entirely.

The companies surveyed hail from Austria, Denmark, Finland, France, Germany, Ireland, Italy, the Netherlands, Norway, Portugal, Spain, Sweden and the UK – together accounting for nearly 90% of EU GDP. Of these, the UK is home to more than a third of AI start-ups in Europe, the report states.

In the wake of the world’s largest consumer electronics show in January, we ran the headline, “AI at CES 2019 – approach with extreme caution” – and Faultline Online Reporter will likely be issuing these warnings for years to come because, unfortunately, this week’s findings won’t change a thing.