(C) Daily Kos
This story was originally published by Daily Kos and is unaltered.
. . . . . . . . . .



How AI Will Change Politics - A Reflection for 2024 [1]

['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.']

Date: 2023-12-30

During his annual Q&A with the public, Russian president, Vladimir Putin, was asked about the dangers of AI by an AI-generated version of himself. This encounter dramatizes the dangers of AI. Putin himself initially seemed confused by the appearance of this doppelganger who looked like him and talked in his voice. The dangers of AI have already spurred the European Union to pass the world’s first regulation on AI, the EU AI Act . The Biden Administration has issued an executive order to help the United States manage the risks of AI. It is clear that AI is the biggest trend of 2023, and will be the biggest trend of 2024 and that it will upend every sphere of human life. AI is eating the world, and politics will not be spared. The question is whether the effects of AI on politics will be net positive or net negative. So far, the consensus seems to be that AI will prove harmful to politics, although there are some researchers who believe that AI will prove good for our political lives. Broadly, there seem to be four ways that AI will change politics: it will lead to more deepfakes and with it, a decline in trust in all content; legislative bodies may accept AI testimony; AI may be used in microlegislation; and AI may also be used as a political consultant.

Deepfakes and the Decline in Trust

The most obvious and immediate danger that AI presents is the use of deepfakes (fake videos or audio). Imagine a scenario where a PAC or foreign actor decided to create a deepfake showing their candidate’s opponent in an election doing something salacious. A Democratic candidate, for instance, could be shown at a fancy restaurant joking about how they do not support abortion rights, in stark contrast to their public statements. Accusation is damaging even when a person is proven innocent, and such a deepfake could be used to derail a candidate. I won’t go into too much detail here because this is an idea that’s been around ever since ChatGPT launched. In many ways, it’s the least interesting. There are initiatives to have images marked to show their veracity, such as Adobe’s content authenticity initiative, but, the obvious counterpoint to that is that people often share images and audio, or stories without checking if they are true. Even if people suspect a story is false, they will still share it anyway, simply because it’s juicy. I can’t imagine that human instinct being halted so that billions of people can stop to check if content has been authenticated. What deepfakes will lead to then, is a decline in trust in content. If you can’t be sure that the content is real, your default stance will be to distrust everything. In a way, that’s a good thing, but it's very easy to go from useful skepticism to conspiracy theories.

AI Generated Testimony Accepted by Legislatures

In science, a simple question has raged: “Can AI invent stuff?” Under existing patent law, AI cannot be named as an inventor, although there have been attempts to get AI on patents. We are not far off, though, from legislative bodies or governmental agencies accepting testimony generated by AI. In Massachusetts and the U.S. House of Representatives ChatGPT has already been used to write bills, and “generate committee remarks,” it has been used in “first floor speech written by AI” events. Although these were heavily criticized for being simplistic or closely guided by very specific prompts, as the capabilities of AI grow, it will become more plausible for legislative bodies and governmental agencies to accept testimony that is wholly written by AI, and submitted as such.

AI As Legislator and Promoter of Rent-Seeking Behavior

Although we are far from a world where AI is elected to office, once we accept that AI content can become acceptable to legislative bodies and governmental agencies, we have to accept that AI could be allowed to amend bills. Amendments to bills range from one sentence amendments to one page. This “microlegislation” is often anonymous, and by its nature, it does not reveal who the beneficiaries of the amendment are. As the Affordable Care Act was being written into law, there were numerous amendments made to it that researchers have said were,

“…replete with rent seeking and other private benefits. For example: “include geriatricians in the definition of primary care providers for the purposes of the primary care bonus”; “reinstate reimbursement for dual energy x-ray absorptiometry (DXA) and vertebral fracture assessment (VFA)”; “increase the Medicare payment rate for nurse-midwives for covered services from 65% of the rate that would be paid were a physician performing a service to the full rate”; “adjust the Puerto Rico Medicare inpatient hospital rate to 100% of the national payment rate.” These rewards are clearly directed at narrow populations.”

With AI, this rent-seeking behavior could scale up, and reduce costs for lobbyists. Rather than hiring a team of lawyers to go through thousands of pages of law, lobbyists could simply feed legislation to the AI, and have proposed amendments done. Higher profits, and more effective rent-seeking, would make America an even more unequal place.

AI As Political Advisor

AI has already had a meaningful impact in some fields. People have lost jobs, or have had their earnings reduced, simply because AI can do what they do, faster, and at a greater scale. Right now, that does sometimes mean that the quality is lower, but for a business, that comes with greater profitability. For instance, given a choice between hiring an artist to create an image for a magazine, or getting Midjourney to do it, most people would choose Midjourney: unless you’re aiming for truly great art, Midjourney is as good as the average artist, and it’s faster, and free. There could come a time when political campaigns turn away from campaign consultants in favor of AI consultants. We have already seen AI-enhanced campaign ads, and we know that AI is capable of so much more. According to Eric Wilson, who thinks we need more AI in politics,

“Democrats have invested in a company called Quiller that helps write fundraising emails and they’ve got a number of candidates doing that. There’s another platform called Localist that’s nonpartisan, that helps campaigns with copywriting.”

Elon Musk’s new xAI could become the Republicans’ go-to AI firm. These initiatives have arisen because tools such as ChatGPT, and Llama impose restrictions on political content. People like Wilson, and Musk and firms like Quiller and Localist, feel that AI should be used in politics.

AI can absorb all the available polling, conduct sentiment analysis on social media, and newspapers, and come up with effective campaign messaging. Campaigns would benefit: lower costs, faster delivery, and output that more effectively reflects the data rather than the biases and leanings of the consultant. Cash-strapped campaigns could have at their disposal a world class consultant, except, it would be AI.

[END]
---
[1] Url: https://www.dailykos.com/stories/2023/12/30/2214554/-How-AI-Will-Change-Politics-A-Reflection-for-2021?pm_campaign=front_page&pm_source=more_community&pm_medium=web

Published and (C) by Daily Kos
Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/dailykos/