Diligent observers of technology markets will realize we have another bubble in the making – the AI bubble. This is not a dismissive statement, AI will change all our lives – for better or worse. But in true style the developers of these technologies, their investors, and users of the technology, will overestimate the short term impact and underestimate the long term impact. We can draw a number of lessons from the dotcom bubble, which burst around 2001.
Prior to the dotcom bubble bursting large numbers of startups were getting funding for some truly silly ideas. I remember one presentation for a business promoting virtual pubs, where people would be home alone, but somehow socialize over the internet. It didn’t happen (thank goodness) and the business disappeared. Technology bubbles are ideal opportunities for the investment community to make a good deal of money, and so greed can overcome good judgement at such times. The bursting of the dotcom bubble saw many startups disappear overnight, and with them the hopes and dreams of millions of investors. But no one would say the Internet has had little impact. It was the perfect vehicle for businesses around the world to exploit self-service in a way that was never previously dreamed of. Customers would enter all their own transaction data using online forms when ordering from a supplier, instead of someone at the end of a phone having to do it for them. Customers would even provide information on new features they would like to see in a product (free market research), and of course the whole thing became electronic with corresponding efficiency improvements. And getting customers to do the work is still an evolving goal. This is what the Internet has delivered – the largest boost to self-service businesses have ever experienced.
When it comes to AI we are just revving up. Dozens of books have been written about the potential impact, and companies like Google and IBM are showing off with self-driving cars and a computer that can win the best players at the game of Jeopardy, respectively. Hundreds of new startups are getting funding, many competing in the same areas, with the hope that one day they will either be acquired or IPO’d. One example is chatbot technology. It seems we’ve all grown tired of seeing other people’s cats, kids and holidays on social media and are using messaging much more. In fact messaging has overtaken social media as the primary way people communicate. It’s more direct and purposeful. Chatbots are automated agents that can carry out a text conversation via a messaging app. Dozens of startups are entering this space, and it’s a good application of AI technology. However there is a strange thing that always happens in information technology markets. Every market eventually becomes dominated by just two or three players. This is true of web browsers, social networking platforms, or anything else you might think of. And so, many of these startups will fail, or possibly be acquired.
Short term thinking is usually preoccupied with novelty. Long term thinking is rare, but in the case of AI we will inevitably see many jobs replaced by intelligent agents – or at least made more efficient by them, requiring fewer people to do the same work. Just as the Internet delivered a platform to dramatically reduce transaction costs for businesses of all sizes, so AI will deliver a platform for businesses to automate much of the decision making. AI is already being used in law firms to automate research, and it is also used in sales situations to identify the best prospects with recommendations of the next best action to close a deal. Just as the Internet allowed business to handle very large volumes of transactions with just a few people, so AI will have the same effect on the decision making processes within businesses. The major impact of AI within businesses will be decision automation. Unlike the Internet, which mainly affected clerical workers, AI will affect all of us – even people who write articles on the topic. Platforms already exist which will scour the Internet and compose an article on whichever topic you desire. So chatbots are already being positioned as a technology that could replace people on help desks, in customer support, and any situation where customers might want dialogue with a supplier.
It is well known that the business use of technology contains an inherent contradiction. As technology displaces human labor, so the distribution of wealth becomes more skewed, and the people who bought stuff have less money to do that. When the sixty richest people in the world collectively have the same wealth as the poorest 3.5 billion, then even those in deepest denial must ask where the customers will come from. This is a trend that seems unlikely to reverse.
Not only will existing businesses, of all sizes, realize massive productivity gains from AI, but entirely new industries will be created – but alas they will not employ many people. The result is likely to be a frenzied use of AI technology to attract the remaining population with disposable income. If you are truly tired of adverts and promotions appearing in your email, social platforms, and now messaging apps – you ain’t seen nothing yet.