Tech entrepreneur and AI expert Shawn Kemp to discuss AI’s potential impact on the upcoming U.S. election and the risks of misinformation, in speaker series event.
On a bright, early-fall day in Boulevard Park in Bellingham, Shawn Kemp poised a profound question, “What does it mean to be human in the age of artificial intelligence?” His voice was low and contemplative, hopeful and curious. Around us, humans went about their business — chatting, walking their dogs, drinking coffee. “What do we give up, and what do we want to keep?”
Kemp, a tech entrepreneur, innovator and artist, will challenge his audience to reflect on this question during his talk at Village Books Sept. 27 titled “The Influence of AI in the Upcoming Election,” part of the Current Affairs speaker series cosponsored by the Salish Current and Village Books.
“That question has been something we’ve grappled with as we’ve gone through different technological advances — the automobile, industrialization, agriculture,” Kemp said. “And I think AI will be a big one.”
AI’s role in politics
Since OpenAI launched ChatGPT in late 2022, artificial intelligence (AI) has boomed and is already affecting aspects of American life. While it has the potential to positively transform industries from health care to energy, the misuse of AI could worsen societal problems by reinforcing biases and spreading misinformation. And irresponsible use in the 2024 election could have monumental consequences.
While AI hasn’t appeared to drastically affect the election or voter behavior so far, a bombshell deepfake — an AI-generated image, video or audio that falsely represents a candidate or undermines their reputation — could still be on the horizon. Kemp, who has worked with AI in some of its earliest forms, including in politics, is poised to deliver a thought-provoking discussion on the topic.
Before his current work as information systems renewal officer at the Non-GMO Project, Kemp started a company called ActionSprout that helped nonprofits and political candidates increase user engagement on Facebook. Part of his work involved using natural language processing and neural networks (precursor AI technology) to analyze Facebook comments from users engaged with certain issues and draw connections between those issues. Imagine a highly complex version of the algorithm that connects artists and songs you might like on Spotify based on what you already listen to.
Using these connections organizations and candidates could form more effective coalitions and campaigns that reached the people who were most engaged or most likely to engage with their cause or campaign.
“Neural networks were able to pull supporters together in ways that we didn’t even realize before, and then we built coalitions and networks around those people to affect change,” Kemp said. “It was pretty powerful.”
Now, AI tools are even more powerful. They can quickly analyze individuals’ social media engagement using similar methods and create highly targeted content most likely to resonate with those individuals.
Deepfakes: a growing threat
But most concerning to Kemp are potential deepfakes that could pull the electorate toward one candidate or another.
Earlier this year, an AI-generated robocall mimicking Joe Biden’s voice urged Democrats in New Hampshire not to vote in the state’s primary.
More recently on Truth Social, Donald Trump posted AI-generated images of Taylor Swift fans — and Swift herself — endorsing Trump. Besides damaging candidates’ reputations, AI deepfakes could also sow doubt in the democratic process if used against election officials or administrators on election day.
Deepfakes are a concern, Kemp said, because they feed our confirmation bias — our tendency to look for information and interpret that information in a way that confirms our own beliefs. Fake news and altered images have exploded across social media platforms and news sites in recent years, but what’s new about AI deepfakes is the speed and ease at which they can be created. In seconds, even amateur AI users can produce false images, videos and voice recordings that, once spread, can strengthen a story people want to believe in.
“It doesn’t matter how fake it is. As long as it feeds the confirmation bias, it has an effect,” Kemp said. “Even if the person believes it was fake, it moves them further down their confirmation bias.”
Preventing AI-generated misinformation
States have begun to pass legislation to regulate the use of deepfakes in elections. Washington, for example, requires clear disclosure when AI-generated audio or visual media are used in election campaigns. Other states have passed outright bans. Federal legislation has been introduced in Congress, but it’s unclear if any will pass before November’s election.
As regulation catches up, Kemp believes our best defense against AI-generated misinformation is not only questioning the validity of content but also thinking more broadly about the process that went into creating it. We should be especially skeptical when we’re fed election-related content that could be trying to sway our vote. Kemp said we need to ask ourselves, “What is the relationship between machine and human” in what we’re seeing?
As the relationship between machine and human increasingly intertwines, Kemp’s question about what it means to be human amid the AI boom feels more urgent. And in the context of politics, Kemp added, if AI is used to craft entire platforms to appeal to the masses, who really is that candidate or political leader? While he doesn’t have the answers to these questions, Kemp believes asking them is crucial for shaping who we want to be in this next age of innovation.
Shawn Kemp explores the intricacies of AI and misinformation in the Current Affairs speaker series, “The Influence of AI in the Upcoming Election,” on Sept. 27 at Village Books, Bellingham, 6 p.m. Reservations are required.
— By Rena Kingery
Also read in Salish Current: “Assistant … or replacement? AI, in real life,” May 12, 2023
An avatar serves as the faculty portrait for a Western Washington University finance and marketing professor. It’s just one small way that both text and images from AI—the subject of both modern news stories and classic science fiction tales—are finding more and more use in the everyday lives of humans.