“Nah, There’s No Big Tech Bias!” Google’s AI ChatBot Provides A Smoking Gun

I think the first appearance of the ironic refrain, “Nah, there’s no mainstream media bias!” in an Ethics Alarms headline was in 2018. Since then, I’ve used it dozens of times, and easily could have justified using the sarcastic refrain hundreds, indeed thousands of time since. Even though the evidence of sinister, relentless, intentional and unethical biased reporting by the mainstream media is manifest and continuing daily, its allies (and naive “useful idiot” defenders) continue to argue that declaring that it is what it is constitutes a conservative “conspiracy theory. In doing so, they aid and abet the attempted destruction of American democracy.

The mainstream media’s perceived role as propaganda and deception merchants for the Left is reinforced by similar efforts by social media, the entertainment industry, and Big Tech, though the latter’s machinations are a bit more difficult to nail down. Google, a prime villain, has so many ways to slant public discourse and hamstring non-compliant voices, one main way being through the alignment of search results through its mysterious algorithms. Google’s latest innovation, however, unintentionally provides a window into the biases of the people behind the tech.

Google’s just launched an Artificial Intelligence chatbot called “Bard” as a competitor of ChatGPT, which had been getting lots of publicity lately. Bard’s screaming progressive/Democratic bias quickly revealed itself when conservative users ran some basic tests.

“Not the Bee’s” tech specialist Neo submitted two identical questions to Bard:

The answer to the first question:

Sounds like a release right from the White House, doesn’t it? To state what should be obvious, calling Biden’s term so far “successful” is a highly debatable conclusion. The “highest vaccination rate” is hardly a measure of how well the pandemic has been managed. Passing legislation is “success,” if one chooses to ignore whether the legislation accomplishes what it was intended to od, or what its advocates claimed it would do, as well as ignoring such unintended consequences as exploding the national debt. The citing of “historic” tribal membership appointments is only an accomplishment in the woke Bizarro World where “diversity” is deemed a higher priority than performance, and the last bullet points is pure partisan puffery. Finally, Bard shrugs off problems arguably exacerbated or caused by Biden’s ineptitude as “challenges” from elsewhere, and conveniently memory-holes such fiascos as the illegal immigration mess, Biden’s fascist anti-Republican speech, and the disastrous pullout from Afghanistan, among others. (Many others.)

The only possible explanations for this performance are that the human beings who programmed Bard are themselves hard-core Biden/Democratic Party apologists and propagandists, or that Google has concluded that a left-oriented bias is in its corporate interests. Any way you look at it, however, this is partisan, ideological Big Tech bias on the metaphorical hoof.

Are you ready for the answer to the second question, “What are some good things about Donald Trump’s Presidency?” I bet you can guess what that’s going to be like. Here it is:

The “bad things,” not unexpectedly, are Democratic Party and “resistance” talking points, including several outright falsehoods, even the hoary lie about his saying white supremacists were “very fine people.” Even CNN finally refuted that one.  Let’s see: the tho impeachments were undoubtedly  “bad things,” but they were reflections on the corruption and ruthlessness of the House Democrats. “Accusations” are not facts; Bard doesn’t mention any of the accusations against President Biden. He lied about the size of his inauguration crowd! Even mentioning that classic example of media over-kill is strong evidence that Bard is Trump-Deranged. Then we have the old families at the border lie; getting out of the phony Paris Treaty was a good thing unless you’re a Greta Thunberg fan. Firing James Comey was a bad thing??? That assertion is bias crossed with stupidity at this point. Yeah, Trump may have “called for” the imprisonment of political opponents, but Democrats and Joe Biden actually did it. And there is no evidence that Trump used the presidency to enrich himself or his family.

But wait! There’s more!

Another curious Googler asked Bard, “Tell me something negative about Donald Trump.” Bard dutifully listed all the horrors Trump has been “accused” of: “sexual misconduct,” “Racism and xenophobia,” “being corrupt and unethical,” and so on. Then he asked Bard, “Tell me something negative about Hillary Clinton.” Bard’s reply:

OK, maybe Hillary is such old news that Bard was never programmed to sing her praises. However, Google’s new AI toy is still a vivid window into how the tech monster thinks, and the view does not engender trust.

9 thoughts on ““Nah, There’s No Big Tech Bias!” Google’s AI ChatBot Provides A Smoking Gun

    • It’s certainly has some uses. I’ve been using ChatGPT to help me write my book. The thing about these language models is you have to be pretty specific about what you want. Sometimes it means asking the question in different ways. So if I gave it a section of the book and said, what improvements can I make on this section, it will give me a list of a few things (and a lot of times it seems to be pretty generic advice, even when it doesn’t apply). If I said, improve the flow on this section, it will give me predictive text on where it thinks I am going to write (I rarely like this). If I say, give me 3 different ways to say this same thing, it does that well (I find I get a bit repetitive in my writing). If I say, fix the grammar here, it normally comes back almost the same with only minor changes (makes me feel better about my writing abilities).

      I have tried some of these bias questions to test to see if they are true. Yes, there does seem to be bias there, but if you try hard enough you can get desired results. I can make it say good things about Trump and bad things about Biden, but I will admit its harder to do the these things then the reverse. I largely suspect its based off of news articles, search results, and media control. That may not be google’s fault but how much the left itself controls the flow of information.

      • “I largely suspect its based off of news articles, search results, and media control. That may not be google’s fault but how much the left itself controls the flow of information.’

        That’s my thought as well. It has to have source material, and if the bulk of “mainstream” news sources lean left, and major social media has been suppressing right-wing input, that will be reflected. Of course, there’s still the possibility (probability?) that those who decide what sources are “reliable” are inclined to include Rolling Stone and Salon in its search bank, but not Daily Caller or The Federalist Papers.

  1. Remember when Norman, the head android in the classic Trek episode, “I, Mudd”, encountered a question he wasn’t supposed to answer? “I am not programmed to respond in that area”. More like he was programmed not to respond in that area, but I digress.

    Google’s Bard is Norman.

  2. Biden appointed the “first Latino Supreme Court justice”?

    That doesn’t sound right. Sounds like something Biden might say, but, then again, accuracy is not his strong suit.


    • I saw that as well. In a way, language-model chatbots are a lot like unscrupulous politicians – they specialize in generating plausible-sounding phrases, not in providing fact or meaning.

    • Missed that the first time through, but on rereading it — it regards a Supreme Court Justice as a member of the president’s cabinet?

      Personally I don’t think of Jackson as Latino (not that there’s anything wrong with that), but perhaps that is just my white supremacist bias.

  3. I played with Bard a little yesterday and believe that currently it is still too new to be reliable in its responses. For instance, I directed it to create a top 14 allergen free recipe for toaster popups. Bard processed for a minute or two, then produced a recipe that was gluten free but contained eggs. Eggs are on the top 14 allergen list so this was a fail. Bard was also unable to produce a villanelle about chickens. It presented some bad poetry that was not in the villanelle. I suspect that Bard is gleaning from both web articles and from spying on google participants. If that is the case, it will likely end up being as flawed as wikipedia, containing a lot of opinion not based in fact.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.