Optimising financial processes

Posted on:

The Emperor’s New Clothes Arrive . . .

I read an article by Tom Davenport in Forbes that absolutely resonated with me.

Titled “Are Boards Kidding Themselves About Generative AI?”, the article reviews a survey of 300 Board members.

An important technological development, with a lot of road still ahead, Generative AI has a host of potential applications.

It is clearly necessary for business leaders to learn about generative AI and be knowledgeable about its likely impacts.

But are they?

Some of the survey responses seem perfectly reasonable, but some beggar belief!

For example, when 76% board members in companies using generative AI were asked about “the collective understanding and knowledge the board of directors in your organization has of generative AI,” they felt they are amazingly well-educated.

28% said the board members are collectively “experts” with a “comprehensive understanding of the underlying technology”.

Another 39% are “advanced,” with “extensive knowledge of generative AI and advanced generative models.”

Just 11% claimed to be “beginners,” and 1% each were “novices” or had “no understanding.”

Hmmmm . . . .

46% of respondents claimed that their Generative AI strategy “is the main priority over anything else.”

Now, THAT’s worrying, maybe enough to make you want to sell their stock!

Tom’s incredulity was reflected in the comment “I am guessing that if the respondents were offered “Generative AI takes the dog for a nice walk,” many would have said yes to it.”

Generative AI is a complex technology in the early stages of application and development.

Trained on billions or trillions of words and using hundreds of billions of parameters to make a decision, we don’t really know how it comes up with a specific response to a particular prompt.

“Hallucinations” (we used to call them “errors”) remain a problem, so we should be very careful about incorporating Generative AI into products and services for customers.

There are plenty of legal concerns.

Even Sam Altman, the CEO of generative AI leader OpenAI and father of ChatGPT, tweeted less than a year ago that “it’s a mistake to be relying on it for anything important right now.”

It is great to be optimistic, but a little balance and humility never goes amiss.

If we think we know everything already, we are a prime example of the Dunning Kruger effect!

Generative AI is unlikely to automate or even transform any “end to end” business cycles or processes, but it does have the potential to make dramatic improvements to certain tasks.

We had to caution the use of the term “Robotic Process Automation” a few years ago, in favour of “Robotic Task Automation”.

Similarly, Generative AI has the potential, especially when developed and trained on curated data in process or task specific “Small Language Models”, to change the game for certain activities that can benefit from a form of “conversational search”.

There is much progress to be made and great opportunities, but we need to be wary of the “Emperor’s new clothes”!

This folk story was a warning to children to think for themselves and be wary of situations in which groups of people praise or seem to believe something of dubious veracity, because they are afraid to go against the general opinion.

Hans Christian Andersen must be turning in his grave!

You can read Tom Davenport’s source article in Forbes here . . .

Thanks for reading . . .