A FLY IN THE OINTMENT

A FLY IN THE OINTMENT

By W H Inmon

Everywhere you look people are going nuts over AI. There are conferences. There are articles. There are books. There are webinars.

Everyone – everyone – has gone crazy over the achievements of AI.

Even the Wall Street Journal seems to have caught AI fever.

Certainly, the venture capitalists have caught the fever. They are dumping millions of dollars into the bet that AI is the wave of the future. And they want to jump on the bandwagon.

But there needs to be a word of caution. Sure, AI is sexy. Sure, AI sparkles and is full of glitter. But let’s take a closer look at what AI really does.

There are many flavors of AI. But at the heart of AI is the ability to use natural language to ask the computer questions. In doing so, AI opens up the computer to the common man/woman.

And that achievement of AI is to be applauded.

But AI operates on a fundamental assumption that simply is not true. That assumption is that the data that AI operates on is available, complete, up to date, accurate and easily understood. And that assumption is simply not true.

Indeed, there is data that can be accessed by AI. But the data that can be accessed by AI is more of a swamp than a mountain meadow. So what good is AI if the data that it operates on is inaccurate or indecipherable.

Let’s take some examples. I create a spreadsheet that says – Bill Inmon makes $1,000,000 per month. Now AI allows the end user to go search for how much Bill Inmon makes. AI quickly finds out that Bill Inmon makes $1,000,000 per month. The problem is that the information that AI has allowed us to access is simply not correct, because Bill Inmon does not make $1,000,000 per month.

As another example, AI allows the end user to ask – how much money did my corporation make this quarter. AI proudly proclaims that the corporation made $12 million this quarter. But AI has retrieved actual revenue, not booked revenue. And then there is sales with forecast revenue. Then there is accounting with their version of revenue. The end user did not know that there are all sorts of different kinds of revenue in the corporation. So, the results returned by AI are questionable, at best.

In the same vein, the end user asks how much money was received. AI goes out and finds US dollars, Canadian dollars, and Australian dollars. AI merely adds the dollar amounts together. Which is fine except that to be added together meaningfully, the Canadian dollars and the Australian dollars have to be recalculated using the current exchange rate.

These simple examples are merely the tip of the tip of the iceberg. The data swamp is much more complex, much more convoluted than these simple examples would have you believe.

AI makes the assumption that there is this pile of data just waiting to be analyzed. It is not a pile of data – it is a garbage dump. And AI is subject to one of the ruling laws of information science – GIGO.

AI is shiny, new, bright. It has been well polished. But AI is building a 20 story building starting at the tenth floor. There is no firm foundation beneath AI.

One wonders if AI will become the tulip mania of this decade.

Bill Inmon lives in Denver with his wife and his two Scotty dogs – Jeb and Lena. Jeb is the rascal and Lena is the little lady. Except that Jeb is teaching Lena all of his bad tricks. Lena now barks when she is hungry and gets agitated by the postman.