There’s far too many discussions going on about what AI can do to improve our world. This isn’t one of those.
This is about what we shouldn’t do. Or, if you’d prefer, why we should think of AI like a mob enforcer.
Like the neighborhood thug, it’s easy to look at the muscle that AI brings to the table as something you’ve got to have. Who doesn’t want the brilliance of AI to digest thousands of pages of research then distill it into a tight one-page summary? But, just like the mob, AI is not content to site on the sidelines waiting to help you. Indeed, AI is already threatening your business.
Interest in the topic has more than doubled in the past five months, according to Google Trends. Most of that is driven by Microsoft’s release of Chat GPT, which has made real for millions what AI can do. This week we have various announcements of the 4th generation of ChatGPT, AI augmentation for Google office tools, and the one everyone has really been demanding, AI for Snapchat.
There’s been some great analysis too, including this paper from the company I work for. It provides a great list of resources if you’re inclined to plunge deeper into the AI world. (Personally, I am intrigued by Do Not Pay, an AI-driven, platform for cranking out disputes, lawsuits and even challenging your tax assessment. And by intrigued, I mean in the way that it’s hard to look away from a car crash.)
So, with that somewhat jaded perspective, let's look again at the state of AI.
Three things to keep you up at night.
Your website sits abandoned, rarely visited – Ask Bing’s AI a question and you get an answer synthesized from millions of web sites that the system has reviewed. What you don’t get is attribution for the facts or even links to the sites, or any branded language that might guide you to the source of information.
The fundamental goal of AI is to make your web site, or at least huge chunks of it, accessible without a click. Your efforts to create compelling content, stunning visuals and a great experience for visitors are worthless to an AI that is just there to extract a fact. Once they have the answers they want, why would a prospective customer come to your site, let alone click on the lead form or sales link?
Pissing off customers when they’re already unhappy – One of the recurring discussions is about hooking up an AI to your company’s chatbot and giving it access to the voluminous reference materials that could help a customer in need. What all of this ignores is emotion. If I walk into your store or call your 800 number (and get through the 13-step decision tree to get to a human) the key component of the interaction will be empathy. We think about these moments as a chance to not only solve the problem, but also tamp down the anger and build loyalty.
Inherent in that successful moment is the voice of your brand, trained and refined through the people you hire and the messages you push out. But how will you instill that into an AI-driven system? Do you have a document that details what the brand voice is, more so than empty words like empathetic, enthusiastic, or informed?
Your company held hostage by anyone with a grudge – Here’s a thought that should terrify anyone in the C-suite. AI tools already on the market enable anyone to produce deep fake videos, photo-quality creations from text prompts, copy in the style of any example, and even voiceovers that are a dead ringer for your campaign. Imagine the damage one pissed-off customer can do to your billion dollars of brand equity with a couple commercials laden with damaging claims, intolerant positions or libelous statements.
Think that’s far-fetched? Consider this Wall Street Journal headline, “As Customer Problems Hit a Record High, More People Seek ‘Revenge.’” It noted that 74% of the 1,000 consumers surveyed said they had experienced a product or service problem in the past year. That is up from 66% in 2020, when the study last was conducted, and 56% in 2017. Those disgruntled customers who have resorted to trying to name and shame their offending companies has tripled in three years.
Need another dose of heartburn? How many small-time competitors would love nothing more than to make a splash by coopting your multi-million-dollar campaign and making it about them instead? Now they can with a couple mouse clicks.
Don't look to the laws
All of this is happening in a space that is, at best, in the margins of established law. Just how long will it take to sort out patent and copyright implications? What is your expectation of privacy if AI has access to everything you’ve ever posted and volumes of data about you? And don’t even try to wrap your head around all the ethical issues.
You don’t need machine learning to realize that it will take legislators and the courts years and years to sort this out. All the while the technology will be evolving at an exponentially faster pace.
All that said, by all means, log in to several AI platforms and play with the tools. Maybe even give it some simple tasks to see how it can help you team. But, under no circumstances, think of AI as a solution to your business’ needs, certainly not anything that currently relies on humans to be successful.
But there is hope
The good news in all of this? We've been here before. When the Internet first became real for most of the world, in the mid-90s, the discussion was full of hype and disillusion. We talked about the promise of a world where everyone is a publisher, and the risk inherent in a world where everyone can say whatever they want.
It has been a while since the Internet flexed its muscle, growing and changing at such a frighteningly fast rate. But, the Internet is a living thing, and with each problem, inevitably, a toolbox of solutions emerges.
The next few years are going to be messy. AI's appeal to our darker side won't go away any time soon. But, with enough thought and effort, we should be able to remind the new thug that we were here long before it arrived. And we'll be here long after it is just a memory. Just remember to keep your eyes open and question everything.
Comentarios