Sorting out the practitioners, pretenders and tricksters in AI

Image credit: pathdoc |

With 40% of “AI” companies not even attempting to use AI in their products and the likelihood of serious disappointment in the short-term, the stage is set for the AI bubble to burst leading to the 3rdAI winter. 

A study by MMC ventures looked at 2,830 European start-ups that have classified themselves as AI companies and found no evidence of any AI in 40% of them. I would take this one step further and run the remaining 60% through RFM’s practitioners vs. pretenders test (see here and here) and would expect to see around 90% of them fall into the pretenders category. This leads me to categorize the European AI landscape as follows:

First, Practioners:

  • These are the companies that are focusing on pushing back the boundaries of what AI is capable of;
  • It is this group that will create sustainable competitive advantages based on the AI that they create that will lead to richer and more intuitive Digital Life services;
  • These are the companies worth paying up for, but I estimate that of the 2,830 companies studied, a maximum of 170 (6%) are likely to be in this category.

Second, Pretenders:

  • These are the companies that are using statistics and data analysis in their products and calling it AI;
  • This does not in any way mean that these are bad companies, but it does mean that the “AI” that they are using will not give them a long-term advantage;
  • Instead, they will have to use other methods such as first-mover advantage or price to create the edge that they need to generate high profit margins;
  • It is on the basis of their success with these methods rather than AI that they should be valued.

Third, Tricksters:

  • These are the companies that say they do AI but where MMC could not find any evidence of AI making a difference to their products;
  • These companies make up 1,132 of the companies surveyed;
  • I think that it is extremely likely that these companies are really just using “AI” to bolster their valuations meaning that in all likelihood, they are meaningfully overvalued compared to their peers. 

MMC data also shows that which is already widely believed which is that companies that are deemed to be involved in AI can command higher valuations when raising money than those that are not. Hence, there is an incentive to make claims on AI regardless of whether one is involved or not.

The net result is more hype, more money being pumped in and higher expectations. The problem occurs when reality has to live up to the hype and RFM research shows that deep learning, in particular, has some serious drawbacks that are likely to prevent this from happening (see here).

The net result will be disappointment, disillusionment, falling valuations and falling investments. These are exactly the characteristics of the first two AI winters, leading to RFM’s prediction that the 3rdis just around the corner.

Winter is coming.

Be the first to comment

What do you think?

This site uses Akismet to reduce spam. Learn how your comment data is processed.