9 Artificial Intelligence Applications That Won’t Work

September 1. 2016. 6 mins read
Table of contents

If you’ve been keeping up with the latest technology trends then you will know that artificial intelligence has the potential to be perhaps the most disruptive technology ever. The bright minds at CB Insights recently put out a graph that shows just how much money is being poured into startups that are “involved in AI”:

Artificial Intelligence FundingOftentimes we get accosted by startups that want us to write about them, and recently we’ve seen an uptick in AI companies that want some publicity. It’s usually the whole “we saw your article on X and was wondering if you could write about our Y technology” type of request. What we will do in all cases is to take a good hard look at whatever it is we’re being asked to write about. Sometimes the companies are interesting enough to merit an article and other times we’re just not that convinced by their value propositions.

Let’s take one financial startup company that claims to be bringing the power of AI and machine learning to the retail investor. They do so by examining “millions of data points per second” and then giving each stock a score of 1 to 10. That’s it guys and gals. They give you a list of stocks and each one has a score of 1 to 10. Just what are you supposed to do with that? We suppose you could setup a trading algorithm that buys 9s and sells 7s which translates stock picking into the crude sort of ranking system that you used to label girls or guys in high school with. This is not a very useful application of AI for retail investors. Going forward, this startup is also looking at chart pattern recognition and also social media data which could be a useful feed for your trading algorithms. However, the usefulness of this site for retail investors that want to apply AI to investing just doesn’t seem to be there yet.

Let’s take another startup which is a content aggregator that breaks down content into smaller pieces. You can then purchase content on a per-article basis as opposed to having to buy a full subscription to the WSJ for example. The startup then claims that they will “use AI to show you content that they think you will be most interested in“. Now first of all, this “one-time purchase” idea could very well just be offered by each publisher but in this case it’s all aggregated in one spot. This is a business model all in itself that doesn’t need AI to make it work. So we signed up for their content and started getting regular emails, however none of the content seemed interesting so we didn’t click any of the articles. Just how is the “artificial intelligence” piece going to figure out what to show us if we’re not interacting with it? Even if we do interact with the AI, there are a finite number of publishers who will adopt the platform so the content set is limited. Unless you are Google News, the usefulness of AI in helping to make content recommendations will be minimal. Forget the AI part for a second and just focus on how you can get the “one-time-purchase” business model to work first.

Let’s face it. While there are a whole slew of legitimate disruptive uses for artificial intelligence like self driving cars and medical diagnostics, there are as many applications of AI that are simply just a technology that’s looking for a problem. We took a stab at jotting down 9 applications for AI that we think won’t be adding meaningful value any time soon.

  • Retail Investing – We’ve talked about this topic before. You can certainly use AI to do some cool things like recognize chart patterns or produce new “big data” sets that you can use for algorithmic trading. The real money to be made here is in having an algorithm tell you to do things that are completely counterintuitive which is what the big hedge funds are doing.
  • Love – Yes, it will only be a matter of time before some dating app claims to “use AI to find your true love”. We’re calling BS on that one. A lot of the people you interact with in online dating are already demonstrating artificial intelligence so that seems tapped out already.
  • Product Recommendations – Amazon has a great recommendations engine that seems to work really well. You’re not going to do this better than Amazon and even if you do, is that marginal difference that you’ll capture using AI really worth the effort? Moreover, will someone pay your startup meaningful amounts of money for it? UDPATE 10/03/2016: A startup called Aspectiva is using AI for product reviews and product recommendations. We stand corrected, and we wrote an entire article on their cool technology.
  • Recruiting – It will only be a matter of time before some recruiting firm decides they’re going to start using AI to “better search the human capital pool to help you win the war on talent”. Don’t buy it. There are a very limited set of white collar workers to match to a very limited set of jobs and it doesn’t take artificial intelligence to do so effectively.
  • Online Marketing – Google Adwords has probably nailed this already, but if you think that you’re going to compete with them using AI then be sure to read about a company called Rocket Fuel. It’s not just about being able to show a better ad, it’s about proving that your solution is better than existing methods, that the AI truly does add value, and that someone will pay for it.
  • Travel Agents – You could have AI come up with your ideal vacation destination taking into account millions of data points about yourself but it’s highly unlikely that this adds any value. If you’re already willing to pay a travel agent to tell you where you ought to go on holiday, then we’ll just tell you that it’s “powered by AI” and you won’t know the difference.
  • Fashion Design – Some fashion company will inevitably decide that the one function in all of humanity that adds the least value to anything, fashion, could somehow be improved by artificial intelligence. You know what? After watching a few fashion shows, we’re going to concede that it IS possible that AI could do fashion better.
  • Car and Home Buying – Yes, we know that your AI powered search engine can find houses and recommend them but there are a finite number of cars and houses and searches are far too constrained for AI to add any value here. Sure, AI could try to predict things like future resale value but let’s face it. You’re not going to commit to making one of the two major purchases in your life based on what AI recommends you should do.
  • Art/Cinema/Music – We just read that Watson put together a movie trailer for a horror movie which is ironically about AI. We have no doubt that AI can do movie editing, create art pieces, and create music. The thing is, there is just no real value add here in having AI do these tasks as opposed to a human aside from the automation aspects. Expect the market to be flooded with all of people’s “artistic AI creations” anytime soon.

Do you remember recently when an artificial intelligence powered software program beat the world champion at Go? The remark was made that the AI player made moves that Go experts had never seen before. Recently IBM‘s Watson was able to diagnose a patient and save her life buy using machine learning. When no other doctor could diagnose this patient, AI could by analyzing and learning from massive amounts of big data. In both these examples AI is doing things that are impossible to do using traditional programming methods.


A good litmus test for an application of artificial intelligence technology is to ask one single question; can this application of AI identify things that we weren’t able to using traditional algorithmic approaches? We fully expect to see loads of startups claiming to be “powered by AI” just to increase their appeal at their next funding round. So what’s the most ludicrous application of AI that you’ve seen yet? Tell us below in the comments.


Leave a Reply

Your email address will not be published.