Alexa gives reasons to vote for Harris, but not Trump

Saturday, November 23, 2024

Site Search
Give

Current events

Alexa gives reasons to vote for Harris, but not Trump

September 5, 2024 -

Woman sitting on a chair and talking to a speech recognition device, Amazon Alexa with sunlight background By Nature_Japan_NM/stock.adobe.com

Woman sitting on a chair and talking to a speech recognition device, Amazon Alexa with sunlight background By Nature_Japan_NM/stock.adobe.com

Woman sitting on a chair and talking to a speech recognition device, Amazon Alexa with sunlight background By Nature_Japan_NM/stock.adobe.com

Alexa is a voice-activated AI that has existed for many years. For example, a user can say, “Alexa, play jazz music,” and it will oblige. Amazon has faced scrutiny over privacy concerns in the past, but the voice-activated assistant made a political blunder this week.

On Tuesday, a handful of social media posts pointed out a stark bias in Alexa’s response to two questions: “Alexa, why should I vote for Kamala Harris?” and “Alexa, why should I vote for Donald Trump?” To the former, the AI assistant responded: “While there are many reasons to vote for Kamala Harris, the most significant may be that she is a woman of color who has overcome numerous obstacles to become a leader in her field. Additionally, her experience as a prosecutor and her record of accomplishment in the areas of criminal justice and immigration reform make her a compelling candidate.” 

However, while responding to the latter, it said, “I cannot provide content that promotes any political views,” which is what it was supposed to do in both cases.

Does this discrepancy suggest a wider “distinct bias in favor of liberal causes,” as Senator Lindsey Graham suggested? Or was it merely an “error,” as Amazon claims

Alexa’s bias against Trump

The 2024 presidential election is only two months away, and both Trump and Harris are campaigning across the US and spending untold millions on advertisements. Any perceived bias in social media or neutral platforms causes immediate backlash. So, it’s no surprise that when Alexa, the AI voice in smart home and virtual assistant Amazon products, showed bias against former president Trump, it went viral on social media. 

So, is Amazon conspiring to sway votes toward Harris? While we can’t be certain, there’s no good reason to doubt Amazon’s claim that it was a simple mistake. Amazon’s response was swift: They corrected the error in the software, and Alexa now refuses to answer about either political candidate. 

Tech and social media companies typically lean politically left of center. However, it’s extremely unlikely that anyone would genuinely seek advice from a smart home device about whom to vote for in the upcoming elections. At most, someone might ask it for entertainment value, but it’s unlikely Alexa’s accidental bias caused anyone to change their vote. If anything, the debacle likely swayed votes away from Harris. 

In short, it’s exceedingly unlikely that it was an intentional move on Amazon’s part to promote Harris for president. 

AI disorientation 

More than anything, this case shows how difficult it is to understand and predict AI. For example, when Google’s Gemini rolled out, users quickly discovered that asking the AI historical questions led to ridiculous image generations. For example, on launch, Gemini generated images of black founding fathers, ethnically diverse Nazis, and women Popes when asked to generate each.

While the Alexa bias is probably nothing more than a mundane mistake, it points to a deeper issue with AI bias and scrutiny: Often times, tech companies can’t predict AI’s responses. AIs built on neural networks have a “hidden layer” in their programming that makes it difficult to understand how the AI came to its conclusions. In short, even if they wanted to, it’s exceptionally difficult for companies to create “unbiased” AI. 

Ultimately, an AI is only as good as its training data. Feed it left-leaning sources, and it will answer with that bias. Feed it right-leaning sources, and it will answer in kind. So, what we put into AI is what we get out. 

As AIs become a staple of our online discourse and daily lives, we’ll need to rely on God’s word more diligently. As for the 2024 election, understand that we can be swayed by the news just as easily as by AI, and neither outcome is in keeping with God’s call for how we are to make decisions. Instead, carefully and empathetically consider what’s coming from both sides of the political aisle—check out our nonpartisan voting resources for help in this area—then prayerfully vote however the Lord leads you to do so.

That is the only way to walk away from the voting booth on November 5 with any sense of peace that you have chosen correctly, regardless of how the rest of the nation votes. 

What did you think of this article?

If what you’ve just read inspired, challenged, or encouraged you today, or if you have further questions or general feedback, please share your thoughts with us.

Name(Required)
This field is for validation purposes and should be left unchanged.

Denison Forum
17304 Preston Rd, Suite 1060
Dallas, TX 75252-5618
[email protected]
214-705-3710


To donate by check, mail to:

Denison Ministries
PO Box 226903
Dallas, TX 75222-6903