Who Ruined AI? (With Tongue Firmly Planted in Cheek)

AI genuinely had a chance to make significant strides to improve the global economy, social justice, health, governance, world peace, and human well-being . . . until it didn’t.

The most recent Gartner Hype Cycle analysis of the status of emerging technologies was published in August 2023. At that time Generative and Augmented AI achieved the dubious honor of holding on (with their fingernails) to the tallest peak of the mountain range of hype. Attaining the peak of “inflated expectations” is an honor to which no self-respecting technology would ever aspire.  The rapid descent into the “valley of disillusionment” is the inevitable bobsled exit from such a lofty honor. Is there any possibility that AI will recover and climb back onto the “plateau of productivity”? Based on similar experiences of other emerging technologies, the odds don’t look good.

Who is responsible for this debacle? How did AI find itself digging its own grave.

There are numerous candidates eligible for the shame and the blame of it all. How did AI hype overwhelm the public’s sense of trust in the technology?

In no particular order of priority, these are some of the culprits now hiding their heads in shame.

Generative AI Explodes on the Scene

ChatGPT, and the spate of other Large Language Models stumbling over each other to be next, successfully mesmerized the world’s population with inspiring dreams of sugar plums dancing in their heads. Before ChatGPT, the topic of AI was a snoozer at dinner parties and cocktail receptions. Only geeks and goons spoke in public about AI applications.

On January 1, 2022, Google Trends data reports that searches about AI were tracking at roughly 10 million per day in the US and UK.  OpenAI’s ChatGPT 3.5 was released by a post on X (formerly known as Twitter) on November 22, 2022. By March 2023, Google searches for AI were running at 100,000 million per day. A tenfold increase in public interest regarding AI over less than four months is rare evidence of truly exponential growth.

Within one week of the Chat GPT release, law firms began advertising their search for “Prompt Engineers” at a compensation of $300,000 annually. When the average US lawyer annual compensation was half that number ($148,000), thousands of law school graduates were scratching their heads relentlessly over how it all could have gone so wrong. Mumbling to themselves in empty law firm corridors (vacant due to the “work from home” mantra of Post-Covid life), associates at every level whispered in tones barely understood by the casual bystander and studiously ignored by partners. Phrases like “College loans, what will I do?” were heard repeatedly as associates careened from wall to wall in a stupor.

All in all, a dire pall was cast over the law firm world of extravagant excess.

Thank you Generative AI.

Big Tech Trembles in Fear

Confronted with the monster to which they gave birth, the self-sovereign Creators of AI gathered to confess their sins to the world. One by one, 350 Dr. Frankensteins of Big Tech sauntered to the altar of shame and threw themselves on the mercy of the world’s population whose very existence was now at risk. On May 30, 2023, four months after the symbolic X tweet and timed to coincide with the testimony of OpenAI’s Sam Altman before Congress, their widely published mea culpa simply but terrifyingly stated:

Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.

Then the billionaire Wunderkinds of tech, jumped back on their Dassault Falcon 8X’s and jetted back to Silicon Valley, Park City, Monaco, or their private islands. A carefully orchestrated public relations blitz was executed to perfection.

It is so satisfying to bare our souls.

Policy Makers In Abstentia

In steady succession, the greatest policy bodies in the world announced their deep concern over the risks of AI and promptly kicked the ball down the halls of regulatory responsibility.

With their fingers held high above their heads to discern the force and direction of political winds, policy makers of the globe’s most powerful nation states have decided to wait to see what the other nation states are going to do.

The EU, the UK, the US, and other democractic superpowers have yet to enact government responses to the “existential threat” of an AI application launched a year ago to public horror and fear. Although compared by AI creators and experts to global pandemic or nuclear war, governmental responses to pandemics and wars tend to be much prompter and decisive than the reaction to Generative AI.

Totalitarian states have responded much more decisively.

Hype Addiction

The “bandwagon effect” from AI has consumed more digital ink than any vaporware in human history. Banks, law firms, publishers, medical care providers, and local donut shops are only a few of the countless business interests that advertise their AI prowess to improve their revenue stream. They encourage customers, clients, patients, and investors to join the waiting list for their upcoming superpower. They entice interested consumers to ask for a demo or “contact us for more information”.

The response? Crickets.

Marketing fluff, website teases, and promises of “more to come” only serve to create jaded consumers. Promising technology yet to be built without delivering on the promises only serves to engender a massive “Bleh!” from the public watching this all too familiar dance.

Failed AI Implementation Projects 

Business leaders often don’t know why, or how, but feel compelled to keep up with the Joneses. The inestimable amount of market buzz about AI leads corporate boards to press their executive teams to come up with the “next new thing”. Executive teams feel obliged to announce their own AI implementation strategy.

The evidence is that up to 90% of AI implementation projects fail to achieve the intended business objectives of the company. It only takes a few costly mistakes to leave executives, business unit managers, and IT teams demoralized and unwilling to risk their careers on the next great idea for AI supremacy.

Fear Mongers

Many skeptics delight in emphasizing and promoting the failures of others. Unbridled enthusiasm leads to irrational hype. However, “Debbie Downers” gain great satisfaction (and unwarranted attention) from irrational pessimism.

Humans are enticed by the extremes of contrary thinking. Find your preferred cognitive perspective on any controversial topic and humans can rarely be dissuaded from vigorously maintaining that position. Confirmatory bias sets in and all we look for is evidence that our opinion is the correct one. What you seek, you will find.

Devaluing the Value of AI

The results are in. Studies of the use of Generative AI have established the fact that users of these text and image generators increase their productive by up to 40% when deployed in their areas of expertise. Taking for granted the accuracy of a Generative AI response outside the user’s domain of knowledge results in 23% less accuracy and value than others performing the same tasks without the use of these tools. Diversity of cognitive thinking in a work group leads to greater innovation, creativity, and productivity. However, a group using Generative AI to help perform their work outside their domain of knowledge experience 41% less diversity of thought.

It is instructive to understand that these new AI applications assist professionals do their work much better when deployed in their areas of expertise, and much worse when they seek wisdom from the AI oracle which they don’t share.

What’s a person to do?

With enough blame to go around, are there reasons to maintain hope in the potential of AI for benefiting humans and advancing societal interests?

At Guardrail Technologies we maintain that AI can be used and managed best when people remain in the loop of engagement and oversight.

The following image depicts the interdependent and evolving process improvement cycle of AI implementation and success.

The path to success with AI applications is the proper preparation of people, processes, and technology to focus the project on its intended goals before implementation of the project begins.

Join us in the pursuit of informed AI applications in this phenomenal age of opportunity. We can deploy AI uses better when we understand what purposes we seek to achieve and prepare the project for success.