The way we seek out and consume information has undergone a seismic shift, and the recent financial tremors felt by one of the world’s largest tech companies offer a stark illustration of this change. When Google’s parent company, Alphabet, reported a staggering $500 billion loss in market value, the culprit was not a failed product launch or a regulatory crackdown, but the very success of its own artificial intelligence. The company’s own earnings revealed that the growing prevalence of AI-generated summaries in search results was directly cannibalizing its traditional advertising revenue. This is because users are finding the answers they need directly in the AI-generated snapshot at the top of the page, eliminating the need to click through to the websites that host the ads. It’s a profound economic signal that the established pathway to knowledge—typing a query and sifting through a list of links—is being rapidly bypassed. The very infrastructure of how we access information is being quietly demolished and rebuilt, with AI as the new, central pillar.
Now, imagine yourself as a voter, on the cusp of an election, trying to grasp the record and character of a candidate. Just a few short years ago, your journey would have taken you through a curated landscape of journalism and collaborative encyclopedias. You might have started with a Wikipedia page, cross-referencing facts with articles from established news outlets, each with its own editorial standards and traditionally verifiable sources. It was a process that, while not perfect, offered a measure of accountability and the opportunity to weigh multiple perspectives. Today, that journey has been radically compressed. A single query now produces an authoritative-sounding, AI-generated synopsis that sits at the top of your screen, distilling billions of data points into a confident paragraph. It preempts any further search. If that distillation feels incomplete, the next instinct is to open a chatbot application and interrogate it directly, posing follow-up questions as if conversing with a omniscient scholar. This new process feels marvelously efficient, but it has traded the transparency of a source list for the opaque curtain of an algorithm. The entire critical thinking process that was once triggered by reading different articles, checking dates, and comparing biases, is often short-circuited by the clean, persuasive prose of a machine that has already done the “thinking” for you.
This new model of information retrieval is precisely what the United Nations warned about two years ago, a warning that has now become a tangible reality in democratic processes from Israel to the United States. The core problem is that while AI answers every question with the unwavering confidence of a seasoned expert, it is deeply susceptible to external influence through a technique known as Generative Engine Optimisation (GEO). It’s the darker, more powerful cousin of Search Engine Optimisation (SEO). In the past, SEO meant gaming the system to get a website to the top of a search results page. GEO is about poisoning the well itself. Malicious actors can use AI to generate a sprawling network of hundreds of fake news websites, flooding the digital sphere with fabricated and distorted information about a candidate. When a genuine AI like ChatGPT or Google’s Gemini goes to learn the latest about a public figure, it scours the live internet and might ingest more of this synthetic slop than legitimate, fact-checked reporting. The AI doesn’t know the difference between a reputable news organization and a fake site; it only sees the weight of available content. Consequently, it molds its responses to a user’s query based on the preponderance of this false data. This tactic is not an abstract theory; it is a sneaky, inescapable, and by no means uncommon trick being deployed in real time. The sheer scale of its impact is staggering when you consider that a significant portion of voters—surveys suggest around 40% in recent U.S. elections—are relying on these AI-generated summaries as their sole source of political information.
The problem extends beyond the words we read to the very images and sounds we see and hear, blurring the lines between reality, fiction, and malicious political strategy. In Israel, the use of AI in political campaigning has become brazenly visual. The Likud party, led by Prime Minister Benjamin Netanyahu, released over 1,800 AI-generated posts during its election campaign. One particularly infamous example depicted a political rival, Gadi Eisenkot, seemingly running past a young man meant to be his late son, killed in the war, in order to embrace an Arab politician. This was a manufactured smear, designed to play on the electorate’s deepest emotions and loyalties. While Israeli law mandates that AI-generated content be labeled as such, this rule is far from global, and its enforcement is sporadic even where it exists. The prolific spread of such deepfakes and synthetic media gives rise to a pernicious phenomenon known as the “liar’s dividend.” This describes a world in which truth becomes so deniable that when a genuine scandal or incriminating video surfaces, the accused can simply dismiss it as another AI-generated fabrication. “That’s not me,” they can claim. “That video is a deepfake.” When the public can no longer trust their eyes and ears, any fabricator can cry foul, and every piece of evidence, no matter how authentic, is cast into a sea of doubt. The result is a total erosion of shared reality, where the very concept of a verifiable fact becomes a casualty of the technological war for power.
The integration of AI into political machinery is not a fringe experiment but a core strategic investment, signaling a future where campaigns are managed by algorithms and bots as much as by human advisors. In the run-up to the U.S. midterm elections, candidates funneled over $130 million into AI-driven content and operations. This investment is not just for the flashy deepfakes; it’s embedded in the mundane but critical work of campaigning. Roughly 80% of campaign strategists now use AI to draft speeches, craft fundraising emails, analyze polling data, and generate personalized answers for voters on social media. This facilitates an unprecedented level of micro-targeting. Voters are bombarded with messages written in their own dialect or language, reflecting their specific local concerns, all crafted by an AI that has digested their digital footprint. In multilingual democracies like the U.S. and India, this is a revolutionary boon for reaching diverse constituencies with pinpoint accuracy. The most concerning application, however, was revealed in Israel, where parties reportedly used AI-powered robocalls to phish voters. These bots, indistinguishable from human canvassers, engage in open-ended conversations, coaxing individuals to reveal their doubts, vulnerabilities, and political leanings. This intelligence is then fed back into the campaign’s data trove, allowing them to tailor future messaging to specifically exploit those expressed anxieties.
In this new world, the challenge for the average citizen has transformed from simply accessing information to actively protecting their own perception of reality. The use of AI in politics and polling is not a passing fad; it is the new norm and will only become more sophisticated, more covert, and more ubiquitous with each passing election cycle. The onus of verification has never been more firmly placed on the individual voter. We must navigate a landscape where the majority of our information might be an AI hallucination, an intentionally generated falsehood, or a micro-targeted piece of psychological manipulation. The old habits of checking sources and questioning headlines are no longer sufficient. We must now question the medium itself, maintaining a deep skepticism for the perfectly crafted, instantaneous answers provided by our AI overlords. To stay connected to reality requires a conscious and exhausting effort to engage with primary sources, to read beyond the AI summary, to listen to human voices in real-time debates, and to intentionally seek out information that doesn’t fit the convenient narratives presented to us. This is the new burden of citizenship—continuously fighting to ensure that our digital tools do not sever our last remaining links to the objective world, and that the authority of the machine does not completely eclipse the judgment of the human mind.