Trusted Sources

I recently created my first AI agent.

Don’t be impressed. It took ten minutes and no technical skill at all. The AI basically synthesizes my news feed.

But maybe I should look into developing an AI chatbot that can convert hardcore right-wingers into progressive activists, because that would be one hell of a cultural improvement from a tool that, so far, has been most effective at writing poorly phrased emails for lazy people and creating bizarro images.

You see, a recent study has found that AI chatbots “can sway voters better than political advertisements.” The researchers found that “a conversation with a chatbot can shift people's political views” more effectively than talking to a human can. But since we’re talking about AI here, and there is always a dark side, you will be unsurprised to know that “the most persuasive models also spread the most misinformation.”

Basically, your Tia Ana will remain steadfast in her political views no matter how many facts you throw at her or regardless of how much you plead. But her principles may waver if an AI bot lies to her enough. 

No, that’s not reassuring.

The researchers say that “the chatbots’ persuasive power could have profound consequences for the future of democracy.” For example, politicians who “use AI chatbots could shape public opinion in ways that compromise voters’ ability to make independent political judgments.”

As we know, many Americans believe Jewish space lasers control the weather and vaccines cause babies to explode. And those beliefs have a head start even before AI does it thing. 

Research shows that the affinity to believe total nonsense is more of a trait for conservatives. This tendency apparently extends to AI, as the study revealed that “chatbots advocating for right-leaning candidates made a larger number of inaccurate claims than those advocating for left-leaning candidates.” 

You see, if AI is parroting right-wing nonsense, it quickly runs out of objective truths. So the chatbots create “misleading or false information” to become more persuasive, and AI “essentially reaches to the bottom of the barrel of stuff they know, so the facts get worse quality.”

The researchers conclude that this study implies “a troubling world for democracy.” But they have a solution. All we need are “strong guardrails to keep the systems in check.”

If AI chatbots could laugh, they would bust out cackling at that statement. Because in the interest of making tech bros even more money, and to facilitate Silicon Valley’s continued right-wing drift, the Trump administration has made it difficult to create any guardrails — never mind strong ones.

Most likely, conservatives will swarm Americans with AI chatbots, deep fakes, and technological tricks designed to fool people into thinking progressives are diabolical and the GOP is competent, neither of which is true.

So in the future, perhaps robots won’t need to rise up, attack humans, and subjugate them. Maybe we will just meekly do whatever AI tells is to do.