AI systems to detect ‘hate speech’ could have ‘disproportionate negative impact’ on African Americans.

If you’re laughing at the headline, you’re a terrible, terrible person.

A new Cornell University study reveals that some artificial intelligence systems created by universities to identify “prejudice” and “hate speech” online might be racially biased themselves and that their implementation could backfire, leading to the over-policing of minority voices online.

Specifically,

[Researcher, Thomas] Davidson said tweets written in “African American English,” or AAE, may be more likely to be considered offensive “due to […] internal biases.” For example, terms such as nigga and bitch are common hate speech “false positives.” “We need to consider whether the linguistic markers we use to identify potentially abusive language may be associated with language used by members of protected categories,” the study’s conclusion states.

“Human error” and “inadequate training” have been cited as explanations.

Update, via the comments:

Given the volume of research that’s subordinate to the conceit that anything reflecting poorly on a Designated Victim Group must therefore, by definition, be an unconscionable act of bias, it’s refreshing to see that the authors of the study do concede that the effect they denounce is most likely a result of statistical differences in actual behaviour:

Different communities have different speech norms, such that a model suitable for one community may discriminate against another… The ‘n-word’… can be extremely racist or quotidian, depending on the speaker… we should not penalise African-Americans for using [it].

However, the authors seem quaintly mystified by the fact that tweets by black people “are classified as containing sexism almost twice as frequently.” And whether the word bitch and various common synonyms should result in flagging and censure only when used by white people and other, as it were, unprotected categories is left to the imagination.

Also, open thread.

Via Neontaster comes news of an intriguing technological development. Apparently, the device “makes the patient feel comfortable,” while “the strong currents impact and rub.” Hey, I’m just reading what it says here.  

Oh, and yes, there is a more intimate video. 

Also, open thread.

Via Darleen, Mary Hudson on the dirty secrets of ‘progressive’ public education:

There was an ethos of hostile resistance. Those who wanted to learn were prevented from doing so. Anyone who “cooperated with the system” was bullied. No homework was done. Students said they couldn’t do it because if textbooks were found in their backpacks, the offending students would be beaten up… I tried everything imaginable to overcome student resistance. Nothing worked. At one point I rearranged the seating to enable the students who wanted to engage to come to the front of the classroom. The principal was informed, and I was reprimanded. This was “discriminatory.” 

Somewhat related, this, this, and the third item here. 

Jonah Goldberg on the immorality of “social justice”: 

Among the myriad problems with this worldview is that individual circumstances are boiled away… Vast abstract categories of human beings are swept up into notions of collective guilt — or victimhood… Traditionally, a person is only supposed to be responsible for the wrongs he or she committed against a specific person. If Person A does something terrible entirely unbeknownst to Person B, it is unjust to hold Person B accountable solely because of the colour of his skin. It’s even more grotesque to hold Person B accountable for the things done by Person A if Person A lived 300 years ago.

Andy Meek on a modern vice: 

Robocallers and spam callers are getting quite good at masking their identity. They do this partly by “spoofing” local numbers, making it seem like a legitimate local number is calling to increase the likelihood that you’ll answer. That’s the reason, according to [caller-ID app] Hiya, that around 9 percent of spam calls a month actually get answered by phone owners even though they don’t recognise the number calling. Nine percent might not sound like much, until you consider the fact that 26.3 billion robocalls were made to American phones in 2018. That’s up 46 percent from 2017’s total of 18 billion.

And via Ace, some news from the world of powerlifting, an activity we don’t often cover here:

Attention, puny humans. The dancefloor is mine: 

Further to the eye-widening James Damore saga of August last year, an update of possible interest: 

James Damore, the Google engineer who was fired after arguing that the gender gap in tech may be partially explained by sex differences among men and women, just filed a class-action lawsuit against Google in the Santa Clara, Calif., Superior Court. Damore came to fame after he wrote the now-infamous “Google manifesto,” where he pointed out that “differences in distributions of traits between men and women may in part explain why we don’t have 50% representation of women in tech.” Filed Monday morning, the class-action lawsuit argues that Google discriminates against white conservative men on the basis of their “male gender” and “Caucasian race,” further alleging that there is “open hostility for conservative thought” in the Google workplace…

The lawsuit also alleges that Google employees were awarded bonuses for arguing against Damore’s views. In one example of this, a female employee was awarded a bonus for speaking out against the “wretched hive of scum and villainy” allegedly represented by Damore’s memo. Email excerpts sent among Google employees after Damore’s memo went public also illustrated pervasive hostility towards conservatives. “I intend to silence these views; they are violently offensive,” wrote senior engineer Colm Buckley in a message to his co-workers. He also wrote to co-workers that there are certain political views “which I do not want people to feel safe to share.”

The details of the lawsuit can be read here and some of the allegations are worth perusing. For instance, those regarding the widespread and enthusiastic use of “shitlists” to exclude and punish employees deemed insufficiently enthused by identity politics; and the apparent indifference of Google’s managers and HR department to harassing, demeaning or racial rhetoric when the targets of such language were white. A phenomenon illustrated by a manager, Chris Busselle, urging white employees – described as “schmucky” and “cheesy” on account of their pallor – to decline invitations to attend conferences; and by another manager, Liz Fong-Jones, who openly professed, in writing, that she “could care less about being unfair to white men.” A view she subsequently described as “absolutely reasonable.”

Update: 

For those so inclined, and via Chris, there’s a crowdfunding option to help out with legal expenses.