AI Circus: Where have the Real experts gone?

 

 "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge." – Daniel J. Boorstin

 

Last year, in a weekly newsletter to their small business customers, my insurance company included a podcast (along with a transcript) titled: What Small Business Owners Need to Know About AI.  Honestly, why would a small business owner or anyone be interested in learning about AI from an insurance company?  Customers may be interested in AI related business insurance products but if they're interested in learning about AI, they'd rather take an AI course, correct? Regardless, it seemed like a good faith effort by the insurance company.

AI, or artificial intelligence, has been one of the hottest topics on almost every CEO, senior and mid-level manager’s agenda since ChatGPT’s birth in November 2022.  Their annual performance appraisals and bonuses probably now also take into consideration the level of AI awareness, training and usage  within the organization.  Not wanting to be left behind, apparently my insurance company had also decided to join this corporate AI bandwagon. What can only be seen as an effort to impress their senior management (think bonus!), the insurance company's editorial team responsible for the content and publication of this newsletter seemed to play bold and approach this podcast with an “Hey small biz guys, guess what? We’re caught up on AI, now let's teach you” attitude.  

The insurance company had engaged the services of an ‘expert’ to create the AI podcast.  According to information available online, this expert had written several books on management, was a columnist at several very illustrious newspapers and magazines, appeared on popular high-profile TV business shows, spoke at various seminars around the country and ran a software implementation consultancy.  With such impressive credentials, it was no surprise that the insurance company believed this expert was best qualified to create the AI podcast.  However, all that glitters is not gold, as I found out about 30 seconds into the podcast.

The AI podcast was broadly categorized into three sections:

(1) Expert's definition of Key AI terms

(2) Expert's description of an AI Policy

(3) Expert's guidance regarding AI Software

Each of these sections is reviewed below with verbatim extracts from the insurance company expert's AI podcast and relevant information from ChatGPT and Gemini as needed.  Closing thoughts follow the section reviews.

 

Expert's definition of Key AI Terms

a. Definition of Artificial Intelligence (AI): The expert started off with defining Artificial Intelligence or AI.  Here’s the expert’s description of AI.  Enjoy.

Insurance Company Expert’s definition of AI

The above definition of AI from the expert is the strangest I’ve heard.  In fact, it sounds more like a snippet from some bedtime horror story.  Or like a warning about some impending alien interference in our routine lives.  

I prompted ChatGPT for a simple definition of Artificial Intelligence.  Here’s the prompt and ChatGPT’s response.

ChatGPT’s definition of AI - Prompt

ChatGPT’s definition of AI - Response

ChatGPT's definition of AI is much clearer and straightforward, wouldn't you agree? Based on ChatGPT's definition, the earliest form of AI technology would be calculators, computer chess (IBM's Deep Blue).  More recent examples of AI include voice assistants like Siri, Alexa and Google Assistant that can understand and respond in spoken language, image recognition such as photo tagging by Facebook, Instagram etc., recommendation systems such as those used by Netflix or Amazon

b. Definition of Generative AI:  Next, the insurance company expert proceeded to define Generative AI.  For your additional entertainment, here are the extracts of the expert’s definition:

Insurance Company Expert’s definition of Generative AI

We’ve surely heard and used ChatGPT a lot by now, but we haven’t quite heard Generative AI defined in the manner the expert did.  So, let’s check out ChatGPT’s definition of Generative AI. 

ChatGPT’s definition of Generative AI

Another straightforward response, courtesy ChatGPT! The answer is in the name: 'Generative' AI i.e. this advanced form of AI generates new content.  It can write articles, stories, create images, videos, audios.

c. Definition of Large Language Model (LLM):  The expert's definition of LLM seemed to be completely out of the left field.  In fact, if LLMs could LMAO then that’s exactly what I’d say the LLMs powering ChatGPT and Gemini did upon reading the expert’s definition.  Here’s the expert’s definition of LLM:

Insurance Company Expert’s definition of Large Language Model (LLM):

No, we’re not confused dear expert, but surely you are.  After failing to provide accurate definitions for AI and Generative AI, you're now struggling to define LLM.  Please note that LLM is not a database as stated above.

Given below are the LLM definitions from ChatGPT and Gemini.

ChatGPT’s definition of Large Language Model (LLM)

Gemini’s definition of Large Language Model (LLM)

 ChatGPT confirmation that Large Language Model (LLM) is not a dataset

 

Expert’s description of an AI Policy

After the inaccurate definitions for key AI terms, the expert made a case for small business owners to “have an AI policy in your business.”  Sure, that made sense.  I continued listening, hoping to hear more meaningful information in this section.  

For the AI policy, however, the expert asked the small business owners to identify the company areas that may need AI and specify the employees who can or cannot use AI.  That’s it.  Nothing else was discussed.  Here’s the extract from the expert’s podcast with regards to the AI policy.

Insurance Company Expert’s description of an AI policy and how to get started

Here’s the issue with the expert’s guidance on the AI policy.  Individual employee access rights may be impacted by an AI policy, but these are not included in an AI policy. Individual access to each application, AI or non-AI, is typically controlled by a company’s IT Security team.  

A policy is generally a higher-level document that provides a set of guidelines and best practices.  An AI policy would therefore provide guidelines and framework for responsible AI use by the company and its employees.  It could address ethical guidelines in AI, company policy around access and usage rights, data privacy, compliance, training, monitoring etc. The policy may identify specific types of AI applications requiring human oversight.  

A small business may choose to combine their policy and access documentation or not have a separate IT security team to handle user security but nonetheless it is important to make this distinction because it is a general best practice.  

Since the expert suggested checking ChatGPT for a free AI policy, here’s an extract of the AI policy guideline from ChatGPT.

ChatGPT’s description of the scope of an AI policy

Expert’s guidance regarding AI Software

Thankfully, the expert reached the final section i.e. the AI Software section, and the rationale behind this podcast started making sense.

Insurance Company Expert’s guidance regarding AI software:

The explanation in this final section seems very straightforward until you realized that this section was the only real takeaway from this entire podcast.

 

 Helping Customers or Networking?

Remember, I’d mentioned earlier that the expert had a software consulting company?  It seems like this podcast was basically just a soft sales pitch for the expert’s software consulting business under the guise of a small business ed on AI.   It gave the expert an opportunity to directly reach a million + users, influencers and decision makers.  This could potentially bring in some new business opportunities for the expert's software consulting company.  

How would the insurance company benefit from this noble learning effort?  The insurance company markets various business-related policies, therefore multiple references to ‘policy’ within the podcast would perhaps gently nudge customers into reviewing their current policies and perhaps upgrading or buying new products? 

Seems like a win-win for both parties. 

 

What did the Small Business Owners gain from the AI podcast?

Confusion – from the bungled up, half-baked definitions and ridiculous explanations of the expert. 

Disappointment – based on the content level and approach of the podcast, it appeared that the insurance company and the expert believed that small business owners were unable to follow complex directions or use advanced technology and wouldn't really understand AI.  Perhaps the small business owners were not an important customer base for the insurance company, therefore their editorial team didn't quality check this podcast prior to releasing it to their million + customers.  An insurance company’s business revolves around assessing, pricing, and managing risk therefore it was quite a bold move for them to risk their reputation over this podcast.

Entertainment – the joke’s on the insurance company and the expert if they believed they were qualified to teach the small business owners or anyone else about AI.  Instead, the insurance company's team should attend some basic AI classes so they can detect the errors in the expert's podcast.  The expert couldn't even define AI correctly, much less create an entire podcast therefore the expert should most certainly take a few AI courses.  Or perhaps the small business owners could teach them a few things about AI?😀

Surprised – when the expert claimed about “speaking a lot” and “writing a lot” about AI. Isn’t anyone vetting this expert?   How do the various world-acclaimed TV, media & publishing companies trust this expert or other such experts to appear on TV shows, write columns, advise their staff, address conferences or provide AI expertise and other software consulting services when they don’t seem to have even the fundamentals in place? 

The AI podcast provided zero value to the Small Business Owners.  

Meanwhile, the expert continues to churn out podcasts for the insurance company on AI and other topics, presumably earning big bucks in the process.  I haven't listened to any of the subsequent podcasts nor do I plan to. These podcasts seem like nothing more than a drain on the insurance company's budget and reputation, especially since the editorial team cannot make an effort to fact check and verify the quality of the final product before releasing it to their customers.  Let's hope someone within the company leadership has the common sense to recognize this wastage and either stop or improve the quality of these podcasts.

Until next time, folks.  Stay sharp, stay curious 🎯🌍✨