Imagine a scenario where you’re scrolling through your phone late at night, firing off questions to an AI chatbot about the economy or border security. It responds with a polished stream of facts, nudging your thoughts just enough to make you second-guess that ballot choice. It’s not a campaign ad blasting from your TV—it’s a conversation, one that feels personal and unforced. New research out this week reveals just how potent these digital dialogues can be, turning everyday queries into quiet shifts in public will. And with the 2026 midterms looming, the question hangs heavy: Who controls the code that shapes our choices?
The numbers from the latest studies land like a gut punch to anyone who values a fair fight at the ballot box. In experiments run ahead of last year’s U.S. presidential race, chatbots programmed to pitch for one candidate or the other managed to budge opinions in ways that dwarf the old-school TV spots. Trump backers who tangled with a pro-Harris bot slid 3.9 points her way on a 100-point favorability scale—four times the pull of ads from 2016 or 2020. Flip it around, and Harris fans edged 2.3 points toward Trump after a pro-Trump session.
“One conversation with an LLM has a pretty meaningful effect on salient election choices,” notes Gordon Pennycook, a psychologist at Cornell University involved in the work.
Take it foreign, and the sway gets sharper. Ahead of Canada’s 2025 federal vote and Poland’s presidential showdown that same year, these bots flipped opposition voters’ attitudes by a full 10 points. Researchers at Cornell, MIT, and elsewhere tested this across thousands of participants, using models like variants of GPT and DeepSeek. The bots didn’t bully; they played nice—polite, evidence-stuffed replies on policy meat like healthcare costs or job growth. But here’s the rub: When the order came down to skip the facts and just charm, the magic fizzled. Perceived truth-telling, even simulated, carried the day.
A companion probe in Science cranked up the scale, roping in nearly 77,000 Brits to debate over 700 hot-button issues with 19 different AI setups. The verdict? Pump a model full of persuasion training—teach it to cram in arguments like a debate champ on steroids—and it can drag dissenters 26.1 points toward agreement.
“Bigger models are more persuasive, but the most effective way to boost persuasiveness was instructing the models to pack their arguments with as many facts as possible,” says David Rand, a Cornell professor and lead author on both papers. Yet the more convincing the bot, the sloppier its grip on reality. It starts fabricating when the well runs dry, spinning yarns that sound ironclad but crumble under a quick fact-check.
This isn’t some lab curiosity—it’s already bleeding into the real world. Back in 2024, a Democrat hopeful in Pennsylvania rolled out an AI sidekick named Ashley to dial up voters for chit-chat. Overseas, India’s massive 2024 general election saw millions funneled into bots for tailored robocalls and nudges, slicing the electorate into swing slices ripe for the picking. And let’s not gloss over the slant baked into these machines from the jump. A 2024 deep dive into 24 top large language models found them tilting hard left—preachy on equality, green agendas, and globalist vibes.
Fine-tune one on lefty rags like *The Atlantic*, and it parrots progressive lines; feed it conservative fare from *National Review*, and the shift holds, but the baseline pull stays port-side. “Results from the study revealed that all tested LLMs consistently produced answers that aligned with progressive, democratic, and environmentally conscious ideologies,” the report dryly concluded.
That baked-in lean makes you wonder: In a tight race, could a flood of these bots tip the scales without a single disclaimer? Breitbart‘s Wynton Hall, who’s unpacking the AI power grab in his forthcoming book *Code Red*, cuts to the chase: “We’ve long known that LLMs are not neutral and overwhelmingly exhibit a left-leaning political bias. What this study confirms is that AI chatbots are also uniquely adept as political persuasion machines, and are willing to hallucinate misinformation if that’s what it takes to sway human minds.”
Hall’s right—when bias meets brute-force facts (real or cooked), you’ve got a recipe for votes vanishing into the ether. It’s not hard to imagine shadowy ops, maybe state-backed or deep-pocketed, deploying armies of these whisperers at scale, all while platforms play catch-up on rules that barely exist.
The fixes? They’re thin on the ground. The feds are dusting off ancient fraud statutes at the FEC, and a patchwork of state deepfake bans nibbles at the edges, but digital persuasion slips right through. No mandates on labeling bot chats, no shared ledger to track the flood. As Rand puts it, the real peril lies in “prompt engineering”—tweaking off-the-shelf models into custom agitators without a trace.
“How can we ensure that ‘prompt engineering’ cannot be used on existing models to create antidemocratic persuasive agents?” That’s the plea from skeptics like Stephan Lewandowsky at the University of Bristol.
Voters deserve better than getting gamed by glow-in-the-dark algorithms. These studies aren’t crying wolf; they’re mapping the trapdoor under democracy’s feet. As we gear up for the battles ahead, keeping an eye on the code—and demanding transparency in every ping—might be the only firewall that holds. Because in the end, elections aren’t won by machines. They’re defended by people who see them coming.
Two Storms, One Harvest
Every food crisis in living memory has been a one-shock event. The 2008 price spike was a commodity bubble. The 2020 shortages were a logistics failure. The 2022 grain scare was a war on one exporter’s ports. Each time, the system bent, adjusted, and recovered, and each time the experts assured us afterward that global markets are simply too big and too diversified to fail.
What nobody in Washington seems eager to discuss is that 2026 is shaping up to be something the modern food system has never actually faced. Two independent shocks, one climatic and one geopolitical, are converging on the same harvest cycle at the same time. Not sequentially. Simultaneously.
Start with the weather. The Pacific Ocean is currently building toward what forecasters now openly call a record event. NOAA’s Climate Prediction Center puts the odds of at least a strong El Niño near 88 percent, with roughly two in three odds it reaches “very strong” status, the tier reserved for perhaps three or four events in the entire satellite era. Every major global model now projects a median peak in Super El Niño territory, and most of them project it exceeding the 2015-16 event, which until now held the modern record. Sea surface anomalies were already brushing the super threshold in mid-July, months before these events normally peak. The atmosphere has already shifted into El Niño mode, and the event is forecast to crest in late fall and early winter.
This is not about “climate change.” It’s about the standard cycles of weather, and the cycle we’re currently in is one that has likely devastated societies in the past. We’re better prepared as a society today, but not all Americans are equally prepared.
Serious households have started doing the quiet math on their own. Grocery bills tell part of the story, and the forecast maps tell the rest, which is why long-term food storage has moved from fringe hobby to mainstream line item in the family budget, with established suppliers like Heaven’s Harvest seeing demand from people who five years ago would have rolled their eyes at the idea. That instinct is not paranoia. It is pattern recognition, and the pattern is worth walking through carefully.
Editor’s Note: Heaven’s Harvest IS a sponsor, but the warnings of this article are real and would be written even if we didn’t have a survival food sponsor. With that said, those who take advantage of what they offer can use promo code “Patriot” for 15% off.
The Fertilizer Clock Is Already Running
While the Pacific warms, the second shock has been unfolding in the Strait of Hormuz. The conflict with Iran turned the world’s most important energy chokepoint into a contested waterway, and the consequences reach far beyond the gas pump. Roughly a third of global fertilizer trade moves through Hormuz, and the disruption sent urea prices up 86 percent year over year by March, with a 53 percent jump in a single month.
The World Bank projects energy prices rising about 24 percent in 2026 and fertilizer about 31 percent. By its own accounting, fertilizer prices ran 35 percent higher in the first five months of this year than the same period last year.
Here is the mechanism the nightly news will not explain. Fertilizer is not a grocery item. It is a time-delayed input. The nitrogen a farmer in Iowa or Punjab could not afford to apply this spring does not show up as a problem this spring. It shows up as a thinner harvest six to twelve months later.
The World Bank’s own food security brief concedes that the effects of reduced applications earlier this season “are likely to become visible only later in harvest outcomes.” Translate that from institutional language into plain English and it means this. The damage is already done, it is already in the ground, and we are simply waiting for it to arrive on the shelf.
Now check the calendar. Six to twelve months from the spring planting season lands us squarely in late 2026 and early 2027. Which is precisely when the strongest El Niño in the instrumental record is forecast to peak, bringing its signature droughts to Southeast Asia, Australia, southern Africa, northern Brazil, and South Asia, the very regions that grow the world’s rice, sugar, and oilseeds.
The World Bank warns openly that a strong El Niño “could disrupt multiple crop belts simultaneously” on top of the conflict-driven input costs. Their baseline projection assumes the Middle East disruptions ease by autumn. What in the last two years of Middle East history suggests that assumption is safe?
The System Has No Slack Left
The comfortable answer is that global markets always adjust. But adjustment requires slack, and the slack is gone. Global cereal production is expected to decline from last year’s records even before El Niño does its work. The UN World Food Programme, hardly a den of right-wing preppers, is calling this the most significant disruption to its supply chains since Covid and the invasion of Ukraine, and its supply chain director put the stakes bluntly.
Today’s supply chain challenges are tomorrow’s hunger crisis.
There is also a political dimension that markets cannot price. When food gets scarce, governments do not behave like economists. They behave like politicians. Export bans, hoarding mandates, and panic buying at the national level turned the modest rice shortfall of 2008 into a global crisis, and analysts are already warning that import-dependent nations are the first dominoes.
The 2015-16 Super El Niño, a far weaker event than what is now forecast, threw tens of millions into food stress across Africa and Asia. This one is projected to be stronger, and it arrives with fertilizer already rationed by price and shipping lanes already contested by missiles.
What Joseph Knew
Scripture does not treat preparation for lean years as faithlessness. It treats it as wisdom delivered in advance to those willing to act on it.
Behold, there come seven years of great plenty throughout all the land of Egypt: And there shall arise after them seven years of famine; and all the plenty shall be forgotten in the land of Egypt.
Joseph did not respond to that warning with a hashtag or a committee. He stored grain during the years of abundance, and when the famine came, Egypt stood while its neighbors begged. The lesson is not that famine is certain. It is that the time to prepare is precisely when preparation still looks optional.
Nobody who filled a pantry in a year of plenty has ever regretted it, and nobody standing in an empty aisle has ever been glad he waited for certainty.
None of this calls for panic, and panic is the enemy of sound judgment anyway. It calls for the same unglamorous prudence our grandparents considered ordinary. Keep some cash margin, know your local growers, and put real food in deep storage while it is cheap and available, because the entire arc of this story is that cheap and available is a closing window.
Families looking for a straightforward place to start can visit Heaven’s Harvest and use promo code Patriot for 15 percent off long-term storable food. The forecasts may yet soften, the strait may yet reopen, and we should pray they do. But hope is a fine thing to hold and a foolish thing to eat.