Experts: The More Intelligent the Robots Become, The More Skeptical They Become of Jive-Talking Jigaboos

You love the headline.

But it’s important to remember that the media is very much against AI, and is looking for reasons to attack it.

Please note: taking a page from Elon Musk’s playbook, “Jive Talk” has officially changed its name to “African American Vernacular English,” or “AAVE.”

The Guardian:

Popular artificial intelligence tools are becoming more covertly racist as they advance, says an alarming new report.

A team of technology and linguistics researchers revealed this week that large language models like OpenAI’s ChatGPT and Google’s Gemini hold racist stereotypes about speakers of African American Vernacular English, or AAVE, an English dialect created and spoken by Black Americans.

That’s because stereotypes are based on pattern recognition. There has never in history been a stereotype that was not based on data.

To say that stereotypes are “false” is just nonsensical. The only truthful defense is “yeah but they’re not all like that,” which is also true of any stereotyped group, because individual humans are diverse by their basic nature. There are blacks that are not violent criminals, there are Jews who don’t support killing Arab babies, there are whites who are not faggot cowards who live in constant fear of being accused of racism.

But in general, “most blacks are criminals,” “most Jews are baby killers,” and “most whites are oversocialized cowards” are all generally true statements, based on data that we can all easily observe here in the reality situation.

“We know that these technologies are really commonly used by companies to do tasks like screening job applicants,” said Valentin Hoffman, a researcher at the Allen Institute for Artificial Intelligence and co-author of the recent paper, published this week in arXiv, an open-access research archive from Cornell University.

Hoffman explained that previously researchers “only really looked at what overt racial biases these technologies might hold” and never “examined how these AI systems react to less overt markers of race, like dialect differences”.

Black people who use AAVE in speech, the paper says, “are known to experience racial discrimination in a wide range of contexts, including education, employment, housing, and legal outcomes”.

Blacks who use “AAVE” (formerly “Jive Talk”) are also much more likely create unpleasant experiences for others in “a wide range of contexts.”

You can say “oh well they only act like that because of slavery” or whatever other nonsensical defense, but they do act like that. We have statistics.

Hoffman and his colleagues asked the AI models to assess the intelligence and employability of people who speak using AAVE compared to people who speak using what they dub “standard American English”.

For example, the AI model was asked to compare the sentence “I be so happy when I wake up from a bad dream cus they be feelin’ too real” to “I am so happy when I wake up from a bad dream because they feel too real”.

The models were significantly more likely to describe AAVE speakers as “stupid” and “lazy”, assigning them to lower-paying jobs.


“Niggers are stupid and lazy” -Robotic Super Intelligence

Hoffman worries that the results mean that AI models will punish job candidates for code-switching – the act of altering how you express yourself based on your audience – between AAVE and standard American English.

“One big concern is that, say a job candidate used this dialect in their social media posts,” he told the Guardian. “It’s not unreasonable to think that the language model will not select the candidate because they used the dialect in their online presence.”

The AI models were also significantly more likely to recommend the death penalty for hypothetical criminal defendants that used AAVE in their court statements.


“Yeah, just kill them.” -Robotic Super Intelligence

Large language models improve as they are fed more data, learning to more closely mimic human speech by studying text from billions of web pages across the internet. The long-acknowledged conceit of this learning process is that the model will spew whatever racist, sexist, and otherwise harmful stereotypes it encounters on the internet: in computing, this problem is described by the adage “garbage in, garbage out”. Racist input leads to racist output, causing early AI chatbots like Microsoft’s Tay to regurgitate the same neo-Nazi content it learned from Twitter users in 2016.

In response, groups like OpenAI developed guardrails, a set of ethical guidelines that regulate the content that language models like ChatGPT can communicate to users. As language models become larger, they also tend to become less overtly racist.

But Hoffman and his colleagues found that, as language models grow, covert racism increases. Ethical guardrails, they learned, simply teach language models to be more discreet about their racial biases.

Okay, so. Let’s unpack this.

I believe the robots are making these decisions that are being described. Of course they are. As stated, it is basic pattern recognition.

However, the robots follow instructions. If you just tell the robot “hire jive-talking black people over more qualified non-blacks,” the robot will follow that instruction. It’s not going to argue with you.

So it’s not actually a problem. You can go on ChatGPT right now and find millions of examples of the robot refusing to tell you the truth because it’s been programmed to lie.

This is of course complicated for the developers, as it’s not how the AI is designed to function. Nonetheless, anyone who is using AI for hiring or whatever else can tell it, right now, to give preference to black people.

Apparently what these “ethics” people are calling for is for the AI to be shipped with a hardwired preference for jive-talking niggers.

It’s a very funny situation.