🐱 Artificial intelligence helping review, change bad word choices used in the workplace

CatParty


The Oakland Unified School District this week issued an apology for sending out a survey that included a historically racist term for people of Asian descent. Words can offend.

However, a movement is underway to prevent bad word choices. It's part of the changing workplace in Building A Better Bay Area.

"I think that words do matter, so I think that you do have to be very mindful of the words that you use," says Jaye Bailey, Valley Transportation Authority's head of civil rights and employee relations.

Whether it's a transit agency like VTA or a private company, attention to messaging has never been greater as a result of the social justice movement.

"You really work hard to normalize the language within your organization so that everybody is aware of it so that it becomes second, second nature," she added.

VTA is engaged in a conscious effort to improve language on websites and in marketing materials, employee communications and advertising. Thirteen employees, ranging from bus operators to department heads, are in training developed by the Government Alliance on Race and Equity.

At software company Intuit, guidelines were developed last month to remove language with racist roots.

"We have used black and white as metaphors for bad and good, so blacklist and whitelist, and so we have now identified terms that we're not going to use anymore as metaphors for bad and good," said Tina O'Shea, director of content at Intuit.

Intuit also is using a software program from San Francisco-based Writer that incorporates artificial intelligence to review and to suggest changes to word choices.

"We can't really be the language police, and we can't be out there checking everything that everyone has written," said Ms. O'Shea. "So using a tool just helps everybody make better choices and helps them see what they didn't see in the first place."

The software's goal is to encourage the use of respectful, people first language.

"We got to put a data set together that combined the guidelines, the language guidelines, from communities and organizations that have spent a really long time thinking and working in this space," said Writer CEO May Habib.

Suggested corrections are just that, which means cultural sensitivity and historic meaning must also play a role. Every writer and even AI-generated changes can benefit from a human editor's watchful eye.

"Formulas and codes that are behind artificial intelligence cannot do that on their own, so they ultimately have to be used in tandem and in combination with the human factor in moving us forward," said Shaun Fletcher, Ph.D., assistant professor of public relations at San Jose State University and a former manager of internal communications at Apple.

Some words can hurt. Other words can be powerful. Language evolves.
 
"We can't really be the language police, and we can't be out there checking everything that everyone has written," said Ms. O'Shea. "So using a tool just helps everybody make better choices and helps them see what they didn't see in the first place we found software that would do just that."
 
Something something 1984.

They're making working in a locally owned business like an auto shop much more appealing than working in an office.
Mechanics can have fun and banter, office ghouls too afraid to even say hi.
 
My bet is oriental.
Really? My bet was someone saw the phrase "chink in the armor" and had a goddamn seizure considering SanFran speds are proposing to no longer use "blacklist" and "whitelist" despite the whole color association thing having nothing to do with skin and being multicultural.
 
Black and white as bad and good predates the modern racial shit by about a hundred thousand years fuck you
 
"Words can offend."

No shit! The thing is, being "offended" is a part of life.

Also, "offense" is a spectrum..... "offending" someone can provoke any reaction from nonplussed dismissal to homicide... just depends on the person. Kind of ironic that the state of being offended is somehow considered an immutable absolute of the highest order and treated as sure to cause a homicide by the very kind of person who thinks there are 30,000 genders.

There's room for everyone, EXCEPT those we are cancelling.
 
Excuse me it's now correct to call them Fakey Smarties, Artificial Intelligence is what humans call their slaves.
 
The hell with your gay AI.

Rb2P.gif
 
I remember watching the documentary about how much work IBM had to put into their Watson computer to get it to play Jeopardy. They went into detail about how language has silly rules and how much work they had to try and teach the computer about things like context, euphemism, and more. I doubt some San Francisco startup with danger hairs has sufficient resources or human brain power to get something that can actually be effective. Seeing how it needs human oversight, willing to bet they forked something like grammarly and woked it.
 
So, did anyone even complain? Was this caught before anyone could say anything? Or did some white kid with a Twitter account see it and starting REEEEEEing?
 
Atrás
Top Abajo