Pinkerton — Hey Siri: How Was the Next Generation of Media Bias Built into A.I.?

The Associated Press
The Associated Press
James P. Pinkerton

There’s a saying in computers: “Garbage in, garbage out.” If there’s bad input, there’s going to get bad output. The same applies to political ideology and computers; specifically, artificial intelligence (AI): liberalism in, liberalism out. 

These thoughts come to mind when the New York Times shares with us that Apple is looking to rev up its AI capacity.  The three-trillion-dollar company, which so excels at making hardware, worries that its voice assistant, Siri, is lagging behind other AIs.  Notably, behind ChatGPT, controlled by Apple’s arch-rival, Microsoft.  ChatGPT’s success has powered the Seattle software giant back into the league-table lead.    

Getting an AI up to speed means inputting more information—and Apple is looking to liberals for the input. The Times explains: “The technology giant has floated multiyear deals worth at least $50 million to license the archives” of news organizations including “Condé Nast, publisher of Vogue and The New Yorker; NBC News; and IAC, which owns People, The Daily Beast and Better Homes and Gardens.” 

Revealing choices.  According to the AllSides news-ratings portal, The New Yorker falls into its left-most category, as it has long “displayed bias through story choice and slant.”  AllSides puts The Daily Beast in the same left category. And as for the fashion magazine Vogue, it slips in plenty of liberal tilt, too. Notably, Melania Trump, as First Lady, was never on its cover, unlike, for instance, Hillary Clinton, Michelle Obama, and Jill Biden. (And as for Teen Vogue, its lefty wokeness is legendary.)

To be sure, not all AI companies have sought to license material; others have simply harvested information wherever they’ve found it.  Yet even this harvesting is revealing: On December 27 the New York Times filed suit against OpenAI over alleged copyright infringement. Thus we are reminded: It’s to the left that AI companies are looking.

Indeed, in its January 8 response to the Times, OpenAI cited its “partnerships” with Main Stream Media outlets, including the Associated Press, as well as, a little more strangely, the overtly left-wing New York University.  The company added, “We are hopeful for a constructive partnership with The New York Times.”  Which raises the question: Are we seeing a new backchannel path for Big Tech to subsidize the MSM?  If MSM outlets get big checks from Big Tech, is that protection money of some kind, or a tacit acknowledgment that they’re on the same side?

The idea that most media leans left is hardly new; and neither is the conservative critique. Back in 1969, a Republican vice president, Spiro Agnew, made hay attacking liberal media bias, setting a tone of red vs. blue oppositionalism that echoes to this day.  Indeed, all new-right media is an implicit critique of old-left media.  That’s the First Amendment in action.  For its part, Apple, which has never hidden its own progressivism, is equally free to use left-wing sources to tutor its AI.  

So by one reckoning, Apple’s choices for liberal media partners are nothing new. Birds of left-feather flock together. Conservatives and populists have been hip to this tendency, and have developed their own counter-organizations.  

Yet now AI comes as a game-changer.  The new technology isn’t just changing media, it’s changing everything. It will transform society and the economy; it will further impact everything we learn and know. Already, there’s abundant evidence that ChatGPT, for instance, leans to the left. That should come as no surprise. The company is headquartered in San Francisco.  

Indeed, the progressive orientation of AI is so deep that it can confound even those sworn to oppose it.  Elon Musk has been attacking the “woke mind virus” for more than a year, and he created his own AI, Grok, to counter the likes of ChatGPT.  Yet just the other day, Grok declared, “Of course, a trans woman is a woman.”  Musk has pledged to fix such excrescences, but in the meantime, we might wonder if there’s a Deep StAIte within Musk’s empire. 

So the non-woke need to be on guard.  Indeed, watchfulness is all the more vital as we move from “hard copy” records.  That is, desk drawers and filing cabinets are giving way to digital files, on hard drives and in the cloud.  That’s all great for convenience and searchability, but anything digital is subject to sudden change.  As in, somebody or something hits “delete,” or alters it in some sneaky way.

Breitbart News has covered the suppression of such un-PC movies as Gone With the Wind, as well as “stealth edits” of classic works of literature, rewritten to meet faculty-lounge standards of political correctness.  Such stealth-editing can be done remotely; a censor in Brooklyn can alter the text of a handheld device in Beaumont or Billings. 

One is reminded of a passage from George Orwell’s 1984:

Every record has been destroyed or falsified, every book rewritten, every picture has been repainted, every statue and street building has been renamed, every date has been altered. And the process is continuing day by day and minute by minute.

Such rewriting and reworking was relatively difficult in Orwell’s “analog” imagination; in these digital days, it’s just an E-Z point-and-click.  And speaking of big brother-ish government, have we considered what the thousands of “disinformation experts” on or near the federal payroll will do if and when they gain full access to databases and archives?  

Moreover, the risk is not just active censorship, but also simple  carelessness and incompetence.  In the 1975 movie Rollerball, set in a dystopic, Great Reset-ish future, a high-tech “librarian” attends the computers that keep all the world’s history.  “It’s embarrassing to misplace things,” the librarian admits.  Misplace what things? “The whole of the 13th century.”  The librarian sighs, “so distracting and annoying! . . . Pity, the 13th century.”  But then he comforts himself: “Not much in the century, just Dante and a few corrupt Popes.” 

So now, if we add AI, we see that knowledge itself is up for grabs.  Information will be free, in the sense that it will be widely available, but only on AI terms.  To cite a practical example, if a college kid wants to “write” a term paper on U.S. history, AI be happy to do it, but the product might be entirely along the lines of the New York Times’ 1619 Project.  

So what to do?  How to stop woke government, and woke corporations, from turning AI into a tool for propaganda and hegemony?  One possible answer is to create an authority to mandate, say, a 21st century “fairness doctrine.”  

But the better answer is what’s always worked for Americans: freedom. Freedom of speech and thought, of course, but also freedom to compete and to innovate. Thirty years ago, it was innovative to see cable news and the internet as right-wing alternatives to newspapers and broadcast TV.  Now we need similar competition and innovation on AI. 

If Apple wants to be progressive, well, that’s its right.  The Constitution protects the left, too.  But the right needs its own AI capacity.  Fight fire with fire.  Musk’s Grok is a start.  More of that, please.  We’ll only be safe when there are substantial tech hubs in Red America. 

The hard work of freedom is many things, from defending the nation to building the economy to forming families.  It’s also working to establish new bulwarks of liberty, including tech structures to protect the sanctity of information and free speech.  The next gener-AI-tion of media bias is already here, and we need to fight that tech fire with counter-fire of our own.  

COMMENTS

Please let us know if you're having issues with commenting.