A.I. — Just a scam or real?
Ever wondered how office-leasing startup WeCompany (formerly known as WeWork) attracted a valuation of $45 billion? — You are not alone. While few people can know the answer for sure, its pitch as an “A.I.-based workplace solution” likely played a big part in it.
WeWork’s anchor investor, the $100 billion Vision Fund, had raised its own fund with an investment thesis around artificial intelligence as “the biggest revolution in human history”. All its 70 investments are positioned as “A.I.-centric”, including WeWork.
In 2014, Wired’s Kevin Kelly proclaimed that “the business plans of the next 10,000 startups are easy to forecast: “Take X and add AI.” As an investor in FinTech companies, I have been on the receiving end of these business plans for the last few years.
It turns out that apparently, 44% of “A.I. startups” don’t actually use A.I. at all, according to this study. Personally, I wouldn’t be surprised if that rate was even higher.
The incentive to dress up as an A.I. company is very real, as “startups claiming to work in A.I. attract between 15 and 50 percent more funding compared to other companies”.
However, startups are by no means the only culprits. Citing “A.I.” is a key theme in the theatre of corporate innovation, and hundred thousands of consultant hours have been billed for A.I. strategies.
Results to date are sobering: A survey of 30 leading banks conducted by the FT revealed that current efforts to apply machine learning are modest, and the problems solved very narrow in scope.
The report quotes the head of Royal Bank of Canada’s A.I. research group:
“The misconception is that humans and machines can perform at the same level. There’s still a long way to go and many challenges we need to solve before a machine can operate at a level even near the human mind”.
Currently, most real-world applications of machine learning are in the domain of chatbots and fraud detection algorithms. Budgets for implementing these applications are limited. Among the 30 leading banks surveyed, they vary from below $3m to $15m. (Contrast these numbers with the $16bn of global VC investment in A.I. startups during 2018).
One of the bank executives participating in the FT survey went on record saying that “it’s very important to separate the marketing story from what the implementation story is, often there is quite a big difference between the two”.
Last week I attended an InsurTech conference in London, and as expected, A.I. and machine learning were the key buzzwords floating around the conference halls. Not surprising, given that few industries rely on A.I.’s foundation — data — as profoundly as the insurance industry. Yet according to consultancy Deloitte, insurers still need to “leap from A.I. mystery to mastery”.
Indeed there are many hurdles for any organization to mastering A.I. — one of them is the shortage of talent. According to professor Julio Diaz Lopez of Imperial College London, “there are 26 million software developers in the world, but only about 300,000 people with the skills to develop AI”.
Apart from the A.I. skills gap, another key challenge for insurers is to make their data machine readable. It is often locked away across hundreds of databases, and generally not enriched with external data. Digital neo insurers such as CommerzVentures portfolio company Getsafe have a head start when it comes to generating data that is usable for machine learning.
A.I. adoption also suffers from security and regulatory challenges. Data privacy regulations such as GDPR and the California Consumer Privacy Act further complicate the effective use of A.I..
The long-term potential of A.I. and its main variants (including machine learning, deep learning, natural language processing, and image recognition) should not be underestimated. We already see first implementations for customer service use cases and in fraud detection.
Indeed, CommerzVentures’ investment in A.I.-based anti-fraud company Fraugster is testament to our belief in the technology’s increasing effectiveness and potential.
However, for the time being, it pays to use a healthy dose of skepticism: The hype around A.I. is far ahead of reality.
