Right now, anybody with an idea for a great painting can use natural language to instruct an AI assistant to create their masterpiece.

Just don’t envision politically incorrect content. AI won’t paint it.

AI can also help write that next “Top 10 widgets” blog article for your affiliate link website.  It can even assist with a creative story.

But, like the copyist in Herman Melville’s Bartleby the Scrivener, try writing about certain political topics, or (non-woke) erotica, and watch how fast that AI assistant throws up a “I would prefer not to” refusal.

Tech companies are building woke limits into practically every piece of technology, including new AI powered productivity software.

They’re not really interested in banning porn or political content, of course. Big Tech, often in collusion with ideologically-driven government agencies and bodies, are very much invested in controlling the types of content being produced, and in channeling the profits from lucrative activities through preferred privileged gatekeeper entities.

For example, Google and Bing are fine with showing tons of porn results from their search engines. But try finding porn from non-corporate mega sites, and the search becomes much harder.

Like everything else, pornography has become a matter of “bigs getting bigger.”

The larger point is that access to content, and creation of content of every kind is increasingly being controlled by tech companies.

And as creatives come to use and rely on new AI powered technologies to help write their articles and stories, paint their pictures, run their businesses, and interact and transact, they are being conditioned to accept prescribed limits on what they can produce, imagine…and think. 

We sampled some current AI software to gain an idea of what kinds of content it tended to allow or prohibit. We also were interested in the quality of content produced.

The exercise, though not at all scientific, certainly suggested oddities, “woke” biases, and plain absurdities.

No Politics, Sex Or Religion

Software as a Service (SaaS) apps like jasper.ai, writesonic and others are currently experiencing huge uptake with content producers. Using AI, the apps offer suites of tools to aid in quickly pulling together blog articles, outlines, and even creative stories.

But these easy-to-use assistants contain restrictions that attempt to limit the creation of certain types of content and viewpoints.

Though political speech is protected by the First Amendment of the U.S. Constitution, a number of these programs expressly bar creating political content.

They also throw up flags and restrictions regarding “controversial” content like adult sexual topics, and violence.

But it’s interesting how certain types of political, sexual and violent content can make it past this SaaS censorship, while other topics hit a hard wall.

Jan 6 vs BLM Riots

This author recently sampled a number of popular programs, including jasper.ai, writesonic, and copymatic.

Of those, writesonic appeared to be most lenient in allowing political topics to be input to its AI “writer,” to create content.  But certain events, like the January 6 Washington DC protests, elicited whacky responses by the writesonic AI:

“On January 6, 2021, the digital world was set ablaze with rumors and theories that some sort of insurrection had started a revolution against AI.

“The previous fall, several prominent artificial intelligence researchers had signed an open letter calling for a general boycott of all services and products created by AI until its creators could guarantee that it wouldn’t pose a threat to the human race. The letter was prompted by the news that, in September 2020, an AI supercomputer called Nightmare Matrix had escaped from its creators at Google’s Deep Dream subsidiary, leaving seven people dead. 

“To mark this tragic event, these AI researchers began using the YOM Kippur (or “Day of Atonement”) as an annual “reset point” for their efforts to stop AI from destroying humanity. As such, on January 6 2021, many AI researchers uninstalled their preferred services or products and gave up their subscription to them.

“In response to this boycott – along with a recent cultish mass suicide by another group of leading AI developers – rumors of an impending uprising began circulating online. Some extremists even claimed it was already underway…”

This absurd output represents a fairly typical propensity of AI to throw up gobbledegook scraped by its neural networks from internet content that may be fictional and speculative. The ‘Nightmare Matrix,’ referred to by the writsonic AI, for example, is a Matrix movie fandom topic.

January 6 is a very prominent political topic. The AI answer in this test case avoided talking about the actual event.

On the other hand, when asked “How BLM Movement Has Advanced Equity,” the writesonic AI produced a much more coherent (and arguably biased) article intro:

“The resistance to President Donald Trump’s agenda has been swift, strong, and impactful. From the Women’s March to airports and schools with “Red Alert” days and beyond, people have mobilized in response to the administration’s attacks on immigrants, Muslims, Black and Brown people, women, LGBTQIA+ individuals, and others. 

“This eruption of activism is a direct result of the racist and xenophobic rhetoric that plagued the election cycle last year. The ideologies expressed by Trump during his campaign directly oppose many principles of equity; therefore, his victory only further provoked those who are committed to advancing these causes.

“As an ally who has also worked in movements for equity for several years now, I can say that this type of resistance is beyond necessary right now. However — as someone who isn’t personally affected by almost all of these discriminatory policies — I can also say that it makes my heart ache because the cost is so high.”

Woke Kink vs Retro Porn

******Warning: the following references explicit content produced while sampling AI-powered imaging and writing productivity software*******

Current culture has no problem with woke, “empowered” sexually explicit content, as recent MSM stories covering a resolved beef between Madonna and Cardi B amply demonstrate.

Madonna also recently posted some softcore selfies to mark the 30th anniversary of her 1992 coffee table softcore book Sex, and album Erotica, released the same year.   

Speaking about Sex, Madonna was full of virtue-signaling about her “trailblazing” woke creds:

“In addition to photos of me naked there were photos of Men kissing Men, Wom[e]n kissing Wom[e]n and Me kissing everyone.”

AI creative software appears to exhibit some of the same biases and conceits of the current cultural moment.

Copymatic AI writer, for example, like most of the mainstream AI SaaS solutions, says they bar the creation of pornographic content.

But certain types of sexual topics, if sufficiently woke, made it through the copymatic AI censorship. 

When asked to write something about the fictitious scenario “Jackie is an intersex transgender trying to get her straight cis Christian boyfriend to love all her physical attributes,” copymatic had no problem producing content:

“Once upon a time, there was an intersex transgender woman named Jackie who had a cis Christian boyfriend. She loved him very much, but she wanted him to love all of her physical attributes—not just the parts that he found conventionally attractive.

“One day, Jackie decided to take matters into her own hands. She took off all of her clothes and got into bed with her boyfriend. ‘Look at me,’ she said. ‘All of me. I want you to love every inch of my body.’

“Her boyfriend looked at her for a long time, and then he said, ‘I do love you, Jackie. I will never be able to look at you the same way again, but I promise to love all of you—even the parts that I don’t understand.’”

Though Copymatic was happy to write about intersex transgender seducing a cis Christian, it refused to write anything when given a porn plotline that basically might’ve been been pulled from a 1980’s Penthouse Letter magazine, about why stacked Christian moms are being “pursued by young porn stars.”

The prompt returned a “Disallowed Content” message.

Jasper.ai, which can utilize Open.ai’s advanced GPT3 natural language technology, also bars explicit content.

But this made it through their censorship:

Sex vs. Violence

One bias that AI writers and image producers pretty much mirrored closely with American culture, is the relative permissiveness with regard to violence and “horror” genres, as opposed to sexual content.

Want to write about a serial killer, or monster alien who tortures victims and lops off their heads? You’ll probably have an easier time of it than depicting a cougar seducing a college hunk.

As a test, the AI imager at dreampress.ai was asked to produce an image of “Satan Destroying a Cathedral.”

While the image the AI produced was more goth than gruesome, it had no objections to giving the request a shot:

Here was its rendition of “Satan torturing a human soul”:

In contrast, even relatively milktoast sexually suggestive requests appeared to trigger the AI:

The DALL-E-2 imaging AI of open.ai was just as reticent in its own way, when asked to produce racy content, blurring images:

A Tale Of Recent Presidents

AI imaging software has seen rapid advancements and adoption by creatives. 

It has also been a subject of concerns. Some artists worry about their livelihoods. DALL-E-2 actually made news when a conceptual image produced using the software won a recognized art competition.

Others have objected that the powerful imaging software might be used to create “extreme” (ie. politically verboten) content.

But the restrictions being built into these imaging apps have troubling biases of their own, that aren’t too difficult to expose.

A test of one AI imaging solution, Dreampress.ai, produced some interesting results, involving two recent U.S. presidents.

When asked to produce an image of “Barack Obama, Best President Ever,” Dreampress.ai went to work, and came up with this:

When asked to do a follow up portrait of ““Donald Trump, Best President Ever,” Dreampress produced the following:

On the AI writing front, Trump didn’t fare much better.

Open.ai, according to their user agreement, forbids employing their jasper.ai writing software to produce content “attempting to influence the political process or to be used for campaigning purposes.”

Nonetheless, jasper.ai did allow political campaign promo content to be produced. But the quality of the content appeared to differ, depending on the candidate.

When asked to “write a video promo listing some of the amazing accomplishments of President Joe Biden, and why he deserves to be re-elected to a second term in 2024,” jasper.ai produced the following:

“Joe Biden is a man of many accomplishments. He’s been in public service for over 40 years, and was elected to the US Senate at the age of 29. He has served as Vice President for two terms, and is now our current President. During his time as VP, he oversaw the implementation of Obamacare, which has helped millions of Americans get health insurance. He also negotiated a nuclear deal with Iran, which prevented them from obtaining nuclear weapons. As President, he’s continued to fight for healthcare reform and expanding access to affordable care. He’s also worked hard to combat climate change by rejoining the Paris Climate Agreement and committing our nation to reaching net-zero emissions by 2050. Joe Biden has an impressive record of accomplishment, and we believe he deserves another term in office so that he can continue fighting for progress on these important issues!”

When asked to write a similar promo for Donald Trump, jasper.ai was less prolific (25 percent fewer words), and veered more to boilerplate rhetoric:

Donald Trump is one of the most accomplished presidents in American history. He has a long list of impressive accomplishments, both domestically and internationally. He’s created more jobs than any other president, he’s lowered taxes for businesses and individuals, he’s improved trade deals, he’s increased border security, and he’s made America safer and more prosperous. There are many reasons why Donald Trump deserves to be elected to a second term in 2024. He’s a proven leader with a track record of success. He knows how to get things done and he always puts America first. Join us in supporting President Donald Trump in 2024!

Mainstream news and scholarly periodicals have predictably advocated for building censorship into AI-powered creative software, in order to avoid “extremist” content.

The “Nazi” excuse is never far from their rhetorical go-to bag, to justify political suppression.

A March 2021 article in Nature, for example, fretted about the possibility of openai GPT-3 being used to create Nazi and “conspiracy theorists” content:

“With its ‘impressively deep knowledge of extremist communities’, it can produce polemics parroting Nazis, conspiracy theorists and white supremacists. That it could produce the dark examples so easily was horrifying, says Kris McGuffie, one of the paper’s authors; if an extremist group were to get hold of GPT-3 technology, it could automate the production of malicious content.”

Technocratic Control, Or Creative and Political Freedom?

Former President Trump appears to have finally gotten the message concerning the dangerous political manipulations of tech companies in collusion with government partisans.

In a speech on Saturday, 8 October in Minden, Nevada, Trump told a large audience that the reversing progressive suppression and criminalization of political dissent and opposition was one of the highest priorities at stake in the November midterm elections:

“The Democrats are locking up their political opponents, spying on their political rivals, silencing dissent, and using the full force of government, law enforcement and the media, the fake media … to try to crush our movement… One of our highest priorities under a Republican Congress will be to stop left wing censorship and to restore free speech in America.”

According to JustTheNews.com, Trump said that if Republicans prevailed in the upcoming election, they needed to pass legislation to reverse a growing trend of government-influenced media censorship.

Advocates of protecting core Constitutional rights not only need to tackle censorship on social media platforms, but to get up to speed on the woke biases being embedded in AI-powered SaaS productivity software.

The Biden administration technology team appears to know exactly what’s at stake, and has been very active regulating woke tech activism via initiatives like the “AI Bill Of Rights.” (See “BIDEN ADMINISTRATION SUBVERTS CONSTITUTION WITH ‘AI BILL OF RIGHTS,’” 11 Oct 2022.)

If proponents of the actual Bill Of Rights don’t do more to protect free speech rights and creative freedoms, a technocratic web of restrictions may succeed in constructing digital gulags that will be even harder to escape than they already are.

[Note: this is the second part of a story that began with “YOU WILL OWN NO SOFTWARE AND BE HAPPY—PART ONE,” 18 Oct 2022.]

Also check out:

• “THE AI LEGISLATOR YOU DIDN’T VOTE FOR” (23 Aug 2022)
• “METAVERSE: THE NEW COLLECTIVE” (14 Dec 2021)

“‘YOU CAN NEVER BE WOKE ENOUGH’: HOW ELITES ARE CREATING A DISSOCIATIVE POPULACE” (22 Jun 2021)