View this post on the web at [link removed]
Last week, consideration of a routine spending bill sparked a political scandal and nearly led to a government crisis. It also demonstrated the tremendous and increasing influence of Elon Musk [ [link removed] ] and his megaphone, X, on legislative processes.
When the fiscal year starts without the necessary funds being appropriated for federal spending, the government can shut down to prevent uncontrolled deficits. To avoid the shutdown, lawmakers have to pass a continuing resolution (CR) bill, which extends spending. The previous CR bill expired on December 20, and a new one was prepared. That’s when the turmoil began.
The urgency to pass the bill was used to make it into a bipartisan deal that included many provisions that have nothing to do with funding the government. Among a number of disparate spending provisions, there were also provisions unrelated to funding at all, such as a provision [ [link removed] ] against deepfake porn, amendments to promote music tourism and a replacement of the term “offenders” with “justice-involved individuals.” The numerous additions have turned the bill into a 1,547-page document [ [link removed] ].
The content of this bill—as well as the practice of urgently packaging unrelated provisions into a spending bill—has caused a lot of criticism, mostly from the right, including President-elect Donald Trump [ [link removed] ] and Elon Musk [ [link removed] ]. Perhaps never before has a bill—not to mention such a massive one—driven such vivid discussion on social media. While legacy media focused on the threat of a government shutdown [ [link removed] ] and the urgent need [ [link removed] ] for essential funding that the bill indeed contained, Elon Musk and his allies made the unrelated provisions in the bill a hot topic on X.
Their strategy seems to have worked. On Thursday, one day before the deadline, the future Department of Government Efficiency, led by Elon Musk and Vivek Ramaswamy, reported on X [ [link removed] ] that the new CR bill contained 116 pages [ [link removed] ] instead of the original 1,547. After intense negotiations, the House passed the new bill on Friday, just hours before the deadline.
This was mostly the influence of Elon Musk and the power of X that brought public attention to the CR bill and affected the legislative process, for better or worse. Yet generally, neither the size nor the language of bureaucratic documents is conducive to being understood or intelligently assessed. While federal legislation is a matter of public record and therefore formally transparent, most of it remains incomprehensible to the broader public.
This will likely change—Elon Musk and Vivek Ramaswamy will use their new megaphone, X, to draw public attention to legislation and policies. They will confront the views prevailing in legacy media, thus providing drama and engagement, as the debate over the CR bill has showcased. However, the issue of the impenetrability of legislative matters and language will remain.
As of now, legislation is so complicated that only experts and lawmakers can decipher and comment on bills with competence. But these interpreters are subject to bias and lack the time or capacity to review every piece of lawmaking proposed in Congress. To combat these problems, author and tech entrepreneur Antonio García Martínez advanced an interesting idea [ [link removed] ] on X: “Consider the political impact of AI ingesting and quickly summarizing long-winded legislation (with all the pork and skulduggery involved) and surfacing all of it to the citizenry. Within seconds/minutes. It’ll be a new form of political reporting.”
Curiously, old media realized a somewhat similar idea 15 years ago—but with the help of crowdsourcing, not AI. In 2009, the U.K. newspaper Telegraph somehow obtained [ [link removed] ] 2 million pages of expense claims submitted by 646 members of Parliament over the previous five years. The paper’s team of investigators dug in and uncovered some sensational facts that led to political scandals and resulted in more than 600,000 [ [link removed] ] additional copy sales.
The Guardian, Telegraph’s major competitor, could have done the same—investigating the huge volume of data in the hopes of uncovering more interesting tips. Instead, the Guardian created a website [ [link removed] ], “Investigate Your MP’s Expenses,” uploaded half a million documents and offered readers the chance to directly check the expenses of MPs.
To attract readers, they made the process fun [ [link removed] ]. The website featured a game-like interface. After viewing a document, a reader could mark it as “Interesting,” “Not interesting,” “Interesting but known” or “Investigate this!” Documents marked for investigation would send a signal to the journalists to investigate further. The interface displayed scans of expenses alongside an MP’s mugshot, and the recognizable faces of their local representatives stimulated voters to dig in and find something. The website featured a progress bar showing how many of the half-million documents had been checked by readers. The front page also displayed a leaderboard showing the top-performing volunteers. The entire project cost 50 pounds ($63) and used open-source software.
Within 80 hours, more than 20,000 volunteers participated, analyzing 170,000 documents. A year later, by June 2010, about half of the 460,000 claim documents had been reviewed [ [link removed] ] by 26,774 registered readers. The project became—and perhaps remains—the largest crowdsourcing initiative in the history of journalism. The Guardian received massive reader engagement, produced a flow of hot stories, led to a full government investigation into MP expenses and earned a reputation as a pioneer in engagement journalism.
It transpired, however, that half the documents on the Guardian’s site remained unchecked. People are lazy and weak; AI is not. Journalism is gone, but AI can dig into legislation bills and reports unwearyingly, missing no detail. All it would take is to feed those 1,547 pages of the continuing resolution bill—or any number of pages of any bill—into an AI. Prompters can ask it to narrate the bill in comprehensible language, find discrepancies and biases, connect it to other documents and policies, make a summary and so on.
There is, of course, the issue of AI bias, as was evident in the recent story about Gemini AI making “diversity errors [ [link removed] ].” An AI reading and analyzing bureaucratic documents and bills for citizens can be twisted to push political or lobbyist agendas. Perhaps even a market of AIs translating bureaucratic documents into human language will emerge. This market will likely mirror the biases of traditional media reporting on government. Multiple diverse AIs, along with the public’s awareness of AI biases, can make this new instrument of political reporting tolerably adequate.
AI as an intermediary between bureaucracy and the people can be many things. It can become a new weapon of political struggle. But it can also make legislation truly transparent, ushering in a new era of governmental openness. One can imagine the shock of bureaucrats and lawmakers when the volume and language of their documents cease to be barriers to public understanding.
Unsubscribe [link removed]?
