Cooking with Frankenstein
I first dipped my toe into the AI stream a couple of years ago and now engage with this hard-to-define new presence every single day.
My current fascination was preceded by dismay. I first tried to use AI to help reconstruct my family’s refugee experience during WWII, which it did with great invention. Worse, some of this invention (called “hallucinations”) made it into Wikipedia, the free crowdsourced online encyclopedia, perverting established historical knowledge.
Two years later, AI is more advanced. I now pay for an upgrade to Gemini Plus — Google’s subscription AI tool. Experienced AI researchers have instructed me to preface every AI query with, “Using and referencing citations for all information presented, what is the (insert your question)?” A young friend calls AI “crack cocaine for the curious.” I have never tried cocaine, but it surely is a heady rush to use such a powerful research tool.
But using AI takes vigilance! As I prepared to smoke some fresh salmon last week, running short on time, I asked AI for sugar-free brine recipes. I was ready to start when some of the AI recipes seemed out of whack, so I asked AI where the recipe had come from. My Frankenstein Chef informed me that it had invented it.
What are the dangers of using AI-invented vs. real human recipes? A page-and-a-half of warnings later, AI summed it up this way: “Blindly trusting an AI-generated recipe can lead to anything from a minor dinner disappointment to a ruined piece of expensive protein — and in rare cases, actual safety hazards … [T]hey don’t actually ‘know’ how ingredients interact in a real kitchen.”
Everywhere I turn, people are learning how to work within the limitations of AI to supercharge their learning, start small businesses and, in endless ways, to indulge their curiosity. But AI is truly a Frankenstein monster, capable of aiding humankind in any endeavor, whether the innocent or the truly vile.
In the hands of those who will steal our goods and — much more frightening, our liberty — it is also crack cocaine for theft.
Corporations and governments are making the largest investments in the history of innovation to develop AI, not to serve individuals’ curiosity or to help me in the kitchen, although that is part of some revenue models. The investment and growth in AI are primarily for darker, baser reasons.
“Calling AI the commodification of almost all human knowledge is not only accurate, but it is also one of the most structurally precise ways to describe what is happening,” Gemini Plus said.
“To commodify something means to take a resource that was once unique, localized or bound to human labor and turn it into a standardized, mass-produced good that can be bought, sold and scaled at a near-zero marginal cost.”
AI is potentially the most powerful tool created since human beings first started using language. But who is going to own it and control it? And for what purpose will they use it? Will a few global corporations use AI to generate unimaginable wealth for a few oligarchs? Will ever-more-powerful totalitarian governments have the ultimate tool for social control?
Or will it be us, ordinary human beings, who use AI to expand knowledge, solve the environmental problems that threaten human survival and unleash the creative and democratic potential of our species?
Gemini Plus doesn’t have an answer for that. No one does.
Peter Pennekamp wrestles with Frankstein at his home in Eureka.
