Futuristic Snake Oil
It’s always the same. New faces appear in the town square. Boisterous voices claiming the impossible. Something about the pitch always sounds downright unbelievable - and the peddlers agree that you’re right to think it sounds just as incredible as your disbelief demands.
But, of course, you’re wrong - because this one is different... This one is industry-leading, clinically-proven, world-renowned, and your reluctance is just a byproduct of your own ignorance. You see, your competition is clamoring, your customers are migrating, your business needs exactly what we’re selling!
What’s old is new again. Generative AI is a technology that is just opaque enough, fanciful enough, to entice via wonder just as easily as it does with fear.
You see, before now we had a consistent trajectory with technology. It was governed by a concept called Moore’s Law. Something that has remained an unwavering constant since the 1960s: The doubling of computer power every two years.
Now? Well, now it’s a little different. Moore’s law has received an exponent on that growth. That doubling previously arriving every two years - it’s now here every six months.
And with that acceleration comes wave after wave of hype and fly-by-night promises.
It’s hard for anyone to keep up on the relevant and legitimate abilities of any new technology, and Generative AI (Gen AI) is eclipsing itself weekly. So, it’s understandable that the current sales pitches for Gen AI apps and tools include a heightened level of urgency along with the blue-sky hand-waving optimism of incredible efficiency gains.
Two opposing, yet symbiotic camps emerge in the SMB x Gen AI space
But first we need to touch on the concept of an API key: It’s basically a handshake between computer programs.
Think of the company you’ve solicited as the one opening the door for you to access another company’s technology. The wrapper placed around the API (Application Programming Interface) is your interface that the third-party service created for you. From there you can adjust and control that foundational technology, Gen AI.
Note: Unique AI-generated phrases are presented as bold and italic.
Now, the first camp consists of third-party services hooked into the APIs of major foundational-model providers. This camp offers quick efficiency gains but often remains hazy on the details.
The second camp comprises SMBs that might be confused or downright fearful of the technology, hesitant to adopt it due to uncertainty and lack of understanding.
These two camps are inadvertently feeding into one another, creating a cycle of hype and apprehension - which then feeds the attention economy, fueling the entrenchment of each camp.
So, let’s break down what’s at the heart of this
There’s a handful of companies with the resources available to train foundational models, these are what we’d call Generative AI.
OpenAI, Google, Microsoft, Meta, Anthropic, Mistral, and a few other major players are those able to piece together the massive compute clusters used to accumulate (and synthesize) the data that is necessary to forge a frontier model. At the time of this writing we could agree this classification is most easily described as ChatGPT.
API keys are available for a monthly fee to third-party businesses to then create tools making use of that foundational technology. The foundational services like Claude, Gemini and ChatGPT’s API allow those businesses to offer any value-added sales pitch that their product-marketing team dares to conjure.
The FTC is light-years behind on figuring out how to regulate any of these claims, and nowhere near ready to require adequate and upfront privacy messaging.
Here’s the issue with that
When all you’re sold is access to a completely different company’s offering, the pitch had better sound downright magical. In fact, some businesses, like Canva, literally call their Gen AI offerings magic!
And if you’re not striking a contract directly with a foundational model provider - OpenAI, Google, et al. - you remain beholden to the same security afforded to all API-key holders. Like the one that any third-party company is using.
Those intermediary companies don’t have the deep pockets that are required to negotiate a Zero Data Retention (ZDR) policy.
Without a ZDR policy in place, this is likely where things would run afoul for your PII and SI handling policies. Don’t allow third-party apps and tools access to any Personally Identifying Information and Sensitive Information of either your business and your customers without clarity on ZDR.
This means do not use any Gen AI tool or app with customer data, sensitive meeting notes, financial documents, or anything you wouldn’t want mistaken as public knowledge about your business or clients.
It’s often necessary to transmit that sensitive business and customer information to deliver on those magical benefits of many sales pitches by third-party Gen AI apps and tools. And that includes companies like Zoom offering meeting summaries. The major enterprise organizations do not use that function, explicitly because of the loose data-retention policies.
I’ve seen upstarts of all sizes offer customer data management tools like CRMs that are supposedly supercharged with Gen AI. …please be cautious when approaching this type of sales pitch.
Allowing third parties access to your data runs afoul of your business’ commitment to the privacy of your customers. It runs afoul of your business’ commitment to itself! Yet this privacy consideration is often buried in the fine print. And we all know how time consuming it is to digest that kind of legalese. Summarizing that policy before signing anything becomes crucially important.
Thankfully, that type of summation is an easy task for Gen AI!
My goal here is to shrink the divide between promise and reality as we navigate through what’s coming next. I want to calm the nerves of those who might be confused about where to begin.
Demystifying the concepts surrounding Gen AI helps reduce fear by providing clear, accurate information. Simultaneously, tempering enthusiasm and watering down the hype ensures that businesses make informed decisions rather than succumbing to unrealistic promises.
By educating SMBs on both the capabilities and limitations of Gen AI, we can foster a more balanced and effective adoption of this technology.
To bridge the gap between the third-party service providers and the apprehensive SMBs, a structured framework for understanding and implementing Gen AI is necessary. This includes clear guidelines on data privacy, realistic expectations of what Gen AI can achieve, and ongoing support to ensure successful integration.
By adopting this balanced approach, businesses can harness the true potential of Gen AI without falling prey to the pitfalls of either blind enthusiasm or crippling fear. This not only promotes sustainable growth but also builds trust and reliability in the evolving landscape of artificial intelligence.
So, for those interested - I offer workshops for SMB owners and their teams to learn the basics and dive into practical applications. I’m also available for one-on-one meetings to talk through what’s possible for your business and industry. Please inquire at cwilke@fellowvector.com, or book time through our website: fellowvector.com.
Note: Unique AI-generated phrases are presented as bold and italic.
The ChatGPT conversation used in the creation of this piece.