Back to Industry Trends & Insights

Getting Real About Exaggerated Claims in the AI Sector

Blog Real

Recent Wall Street Journal Story Discusses Overstated Claims from AI and Tech Firms; How to Determine What is Real

By Steve Butler, Founder & President, AI Foundry

A recent story in The Wall Street Journal, “AI Startup Boom Raises Questions of Exaggerated Tech Savvy” identifies a troubling issue - companies that have exaggerated their AI capabilities in order to attract customers and investors. The article explains that there is a growing challenge to figure out how proficient a company really is in AI, and often there is a gap between where they actually are and where they say they are.

However, at the same time, investment in the AI sector is booming. The story states that venture firms nearly doubled their funding of AI startups in 2018 (to $31 billion) vs. 2017, and that SoftBank recently started an AI-focused investment fund with $108 billion in expected capital. Also, companies with AI in their description have raised 15%-50% more funding than other software startups – so there is considerable temptation to stretch reality when it comes to describing one’s AI capabilities.

The story details how one company inflated its claims to AI technology, with a system that claimed to have cutting-edge AI, but was actually using conventional software and manual labor. This article points to an alarming trend that concerns both customers and investors and could harm the long-term development and reputation of the AI industry.

So, how can you determine what is “real AI” and what is a big exaggeration?

Determining What is Real in AI

What many people don’t know is that there are AI companies out there who don’t work with customers and have not faced real production environments. They might have an algorithm, a theory, or a small test case – and some funding – but they are not working with real data from real customers.

And you can only do so much “synthetically” with data – you need to have real customer data coming at you and be able to process it and work with it the right way. This is why Tesla put cars on the road as soon as they could…they needed to collect real data as soon as possible. You have to be “out in the wild” to get that kind of real customer data (and different types of data) – this is an important building block for developing an effective AI solution.

AI Foundry – A Real AI Platform with Real Customers

At AI Foundry, we have combined several technologies, including the latest in imaging, AI, machine learning, big data and BPM to create the Cognitive Business Automation Platform. We have an advantage in that we have a real AI-based platform, and we have real customers who are using it today to help improve their business. One of those customers, Radius Financial Group, was quoted in a (different) Wall Street Journal story earlier this year.

And, the case studies section of our website includes examples of other customers using our platform.

We already have a strong foundation based on the documents that we can automate now, with AI-based extraction, classification and intelligence capabilities. And, we have a bright future ahead of us based on new things that we will be able to automate soon. Our talented staff of data scientists, engineers and mortgage professionals give us the technical and industry expertise to keep building upon our platform and helping more customers.

The bottom line is that we are attacking a problem that exists today and we are solving it. This makes us a rare but important player in the AI market – one with a real AI product and real customers.