The AI Has No Clothes

The term “Unified Communications” has been around since the early 2000s.  Great, except that’s incorrect. 

I was writing an article on UC&C last week and I forgot the exact date that the term Unified Communications was coined.  I knew it was in the 1990s, but not sure exactly when.  When I asked ChatGPT the question, here was its reply:

My memory on the subject felt that was a completely wrong answer, so I continued to search for the right one.  Wikipedia had the correct answer: 

The term unified communications arose in the mid-1990s, when messaging and real-time communications began to combine. In 1993, ThinkRite (VoiceRite) developed the unified messaging system, POET, for IBM's internal use. It was installed in 55 IBM US Branch Offices for 54,000 employees and integrated with IBM OfficeVision/VM (PROFS) and provided IBMers with one phone number for voicemail, fax, alphanumeric paging and follow-me. POET was in use until 2000.[5] In the late 1990s, a New Zealand-based organization called IPFX developed a commercially available presence product, which let users see the location of colleagues, make decisions on how to contact them, and define how their messages were handled based on their own presence. The first full-featured converged telephony/UC offering was the Nortel Succession MX (Multimedia eXchange) product, which later became known as Nortel Multimedia Communications Server (MCS 5100).

Still concerned about the AI ‘hallucination’ I went back to ChatGPT and quickly realized I could ask it the same question using different phrasing and get a different answer each time.  On top of that, I was recently made aware (by fellow analyst Kevin Kieller ) that I couldn’t ask a large language model like ChatGPT why it gave me one answer over another one, because every time you ask an LLM why it gave you an answer it just responds with what it determines is the most appropriate answer to that new question, not the actual facts around the first answer.  That’s a brain-exploding concept once you really try to understand it (thanks a lot Kevin.)

Which brings me around to the point of this article on AI – we need to calm the heck down with all the AI hype (and heck was not my first choice of words there.)  We spent a few years touting the Metaverse as the next thing that was going to change civilization, and only stopped over-inflating the concept and boasting about it when this next big fad (AI) came along.

Should we all be aware of the concept of AI and how and where to interact with it?  Well…sure…OK…that’s a good thing I suppose.  Do we want to live in a world where large data centers write our articles, poetry, school assignments, emails, advertising campaigns, meeting summaries, legal briefs, etc., etc.? Um, no – that should be obvious to anyone that is still capable of thinking for themselves.  (Regrettably, that is becoming a smaller list every day.)  Both students in school and post-graduate lifelong learners have embraced the idea that AI can do everything for them, and the results of this are already showing the signs of those human skills beginning to be atrophied.  What will we call the next generation of employees in the workforce that can’t effectively think for themselves?  Generation Dumb?  Gen AI?  Gen Dependent?    

In today’s hype-cycle-saturated environment, prestigious companies are telling us that we need to rely upon (and sometimes pay handsomely for) their AI, and set-up our businesses and lives to be dependent upon its support…and just like with the Metaverse fantasy, people are buying into the hype.     

In addition to the frightening concept that we would leave our most critical thinking to faceless data centers that are not even capable of explaining their logic, there is another sword of Damocles, ready to plummet and shatter the AI space.  There are a bunch of lawsuits against AI organizations for stealing intellectual property for training without permission.  Copyrighted newspaper articles and books can clearly be traced as the source of much of the information being served-up by these AI thinking-crutches – a practice that is clearly against the law.  If you thought you’ve lived through some societal bubbles bursting, just wait for this one to pop.  The resulting shock-wave will likely reduce some companies to rubble.

For my entire professional career, I have never been afraid of being the little boy who points and shouts that the emperor has no clothes.  This is – clearly to me at least – another one of those naked emperor moments.  AI and LLMs are wonderful tools – period.  Choosing the correct tool for each job is a human skill, not one we should be leaving to data centers. 

Understand AI to the extent possible? Sure.  Learn how to prompt AI to obtain more accurate answers and less hallucinations? Yes.  But buy-into the idea that you should be embedding AI into your business processes at the current time? Um, no…that’s not just bad, it’s very risky.  Hasn’t your Risk Management team told you that, or have you already laid-off that team in favor of investing your capital into AI?  What capital will you have left when the next fad and hype-cycle comes along?  Moreover, who will buy your products and services when all those employees that were laid-off no longer have the funds to buy them, and the firms that laid them off no longer have the need for all those licenses to use your product or service? 

If there were ever a time for organizations to realize that AI is just one of the proverbial emperor with no clothes moments, this is it.

====================================================

PLEASE click the “Like” button on this blog if you liked it. It will allow others in your network to see it. Also, feel free to comment to join the discussion.

====================================================

This article was written by David Danto and contains solely his own, personal opinions. David has over four decades of experience providing problem solving leadership and innovation in media and unified communications technologies for various firms in the corporate, broadcasting and academic worlds including AT&T, Bloomberg LP, FNN, Morgan Stanley, NYU, Lehman Brothers and JP Morgan Chase. He is also the IMCCA’s Director of Emerging Technology. David can be reached at Perspectives@danto.com or DDanto@imcca.org and his full bio and other blogs and articles can be seen at Danto.info. Please reach-out to David if you would like to discuss how he can help your organization solve problems, develop a future-proof collaboration strategy, or if you would like his help developing solid, user-focused go-to-market strategies for your product or service. Visit the website TechPerspectives.info for more information.