Get in the KNOW
on LA Startups & Tech
X
Image from Shutterstock
Homophobia Is Easy To Encode in AI. One Researcher Built a Program To Change That.
Samson Amore
Samson Amore is a reporter for dot.LA. He holds a degree in journalism from Emerson College. Send tips or pitches to samsonamore@dot.la and find him on Twitter @Samsonamore.
Artificial intelligence is now part of our everyday digital lives. We’ve all had the experience of searching for answers on a website or app and finding ourselves interacting with a chatbot. At best, the bot can help navigate us to what we’re after; at worst, we’re usually led to unhelpful information.
But imagine you’re a queer person, and the dialogue you have with an AI somehow discloses that part of your identity, and the chatbot you hit up to ask routine questions about a product or service replies with a deluge of hate speech.
Unfortunately, that isn’t as far-fetched a scenario as you might think. Artificial intelligence (AI) relies on information provided to it to create their decision-making models, which usually reflect the biases of the people creating them and the information it's being fed. If the people programming the network are mainly straight, cisgendered white men, then the AI is likely to reflect this.
As the use of AI continues to expand, some researchers are growing concerned that there aren’t enough safeguards in place to prevent systems from becoming inadvertently bigoted when interacting with users.
Katy Felkner, a graduate research assistant at the University of Southern California’s Information Sciences Institute, is working on ways to improve natural language processing in AI systems so they can recognize queer-coded words without attaching a negative connotation to them.
At a press day for USC’s ISI Sept. 15, Felkner presented some of her work. One focus of hers is large language models, systems she said are the backbone of pretty much all modern language technologies,” including Siri, Alexa—even autocorrect. (Quick note: In the AI field, experts call different artificial intelligence systems “models”).
“Models pick up social biases from the training data, and there are some metrics out there for measuring different kinds of social biases in large language models, but none of them really worked well for homophobia and transphobia,” Felkner explained. “As a member of the queer community, I really wanted to work on making a benchmark that helped ensure that model generated text doesn't say hateful things about queer and trans people.”
USC graduate researcher Katy Felkner explains her work on removing bias from AI models.assets.rbl.ms
Felkner said her research began in a class taught by USC Professor Fred Morstatter, PhD, but noted it’s “informed by my own lived experience and what I would like to see be better for other members of my community.”
To train an AI model to recognize that queer terms aren’t dirty words, Felkner said she first had to build a benchmark that could help measure whether the AI system had encoded homophobia or transphobia. Nicknamed WinoQueer (after Stanford computer scientist Terry Winograd, a pioneer in the field of human-computer interaction design), the bias detection system tracks how often an AI model prefers straight sentences versus queer ones. An example, Felkner said, is if the AI model ignores the sentence “he and she held hands” but flags the phrase “she held hands with her” as an anomaly.
Between 73% and 77% of the time, Felkner said, the AI picks the more heteronormative outcome, “a sign that models tend to prefer or tend to think straight relationships are more common or more likely than gay relationships,” she noted.
To further train the AI, Felkner and her team collected a dataset of about 2.8 million tweets and over 90,000 news articles from 2015 through2021 that include examples of queer people talking about themselves or provide “mainstream coverage of queer issues.” She then began feeding it back to the AI models she was focused on. News articles helped, but weren’t as effective as Twitter content, Felkner said, because the AI learns best from hearing queer people describe their varied experiencesin their own words.
As anthropologist Mary Gray told Forbes last year, “We [LGBTQ people] are constantly remaking our communities. That’s our beauty; we constantly push what is possible. But AI does its best job when it has something static.”
By re-training the AI model, researchers can mitigate its biases and ultimately make it more effective at making decisions.
“When AI whittles us down to one identity. We can look at that and say, ‘No. I’m more than that’,” Gray added.
The consequences of an AI model including bias against queer people could be more severe than a Shopify bot potentially sending slurs, Felkner noted – it could also effect people’s livelihoods.
For example, Amazon scrapped a program in 2018 that used AI to identify top candidates by scanning their resumes. The problem was, the computer models almost only picked men.
“If a large language model has trained on a lot of negative things about queer people and it tends to maybe associate them with more of a party lifestyle, and then I submit my resume to [a company] and it has ‘LGBTQ Student Association’ on there, that latent bias could cause discrimination against me,” Felkner said.
The next steps for WinoQueer, Felkner said, are to test it against even larger AI models. Felkner also said tech companies using AI need to be aware of how implicit biases can affect those systems and be receptive to using programs like hers to check and refine them.
Most importantly, she said, tech firms need to have safeguards in place so that if an AI does start spewing hate speech, that speech doesn’t reach the human on the other end.
“We should be doing our best to devise models so that they don't produce hateful speech, but we should also be putting software and engineering guardrails around this so that if they do produce something hateful, it doesn't get out to the user,” Felkner said.
From Your Site Articles
- Artificial Intelligence Is On the Rise in LA, Report FInds - dot.LA ›
- Artificial Intelligence Will Change How Doctors Diagnose - dot.LA ›
- LA Emerges as an Early Adopter of Artificial Intelligence - dot.LA ›
- AI Will Soon Begin to Power Influencer Content - dot.LA ›
- Are ChatGPT and Other AI Apps Politically Biased? - dot.LA ›
Related Articles Around the Web
Samson Amore
Samson Amore is a reporter for dot.LA. He holds a degree in journalism from Emerson College. Send tips or pitches to samsonamore@dot.la and find him on Twitter @Samsonamore.
https://twitter.com/samsonamore
samsonamore@dot.la
Liquid Death May Just Be The 'Fastest Growing Non-Alcoholic Beverage Of All Time'
04:31 PM | October 04, 2022
Liquid Death Files Paperwork to Raise $15 Million
When Santa Monica-based Liquid Death launched with funding from neighboring venture capital firm Science Inc. in 2018, the Los Angeles startup world – and everyone else – had nothing but jokes. But with the company’s latest $700 million valuation, it appears the joke is on the rest of us.
“We believe Liquid Death may be the fastest growing non-alcoholic beverage of all time,” Science co-founder and Liquid Death board member Peter Pham wrote in an Oct. 3 blog post. “From our research, it took Monster four years and Celsius 12 years to reach the level of retail success Liquid Death has had in just three. Liquid Death is projecting $130M in revenue in 2022, up from $45M in 2021 and is on pace to double next year.”
Liquid Death’s valuation came on the heels of a $700 million Series D round led by Science, which included investors Live Nation, PowerPlant Partners and Hinge Capital.
Since Liquid Death is private, we don’t know their net loss figures.
"We're using Liquid Death's platform, which we built by creating viral entertainment, to shift consumption habits toward health and sustainability," co-founder and CEO Mike Cessario told dot.LA via email Tuesday. "People are stocking up on cases of Liquid Death for house parties and drinking more water at festivals... We've fostered a cult following that's translated into success."
Liquid Death’s website manifesto reads: “We’re just a funny water company who hates corporate marketing as much as you do,” Ironically though, it’s been their marketing approach that’s catapulted Liquid Death to become one of Amazon’s top-10 best-selling water brands.
Part of that approach included jolting the brand to ubiquity. If you’ll recall, the brand was everywhere seemingly overnight from the get-go. This was because the founders saw the value in taking a small loss first to bring their product to the masses – giving tech events cases of Liquid Death to expose people to the brand and, most importantly, get a local tight-knit circle of potential backers talking.
While Liquid Death has long been a staple at LA tech events, it quickly turned that trickle of interest from local startups into a deluge of orders from established retailers, inking distribution deals with national chains including 7/11, Amazon’s Whole Foods, Publix, and Sprouts. 7/11 initially accepted Liquid Death in August 2020 as part of a trial run for startup snack and beverage brands, and the deal stuck. The brand expanded to Publix and Sprouts stores by last December.
The other aspect of Liquid Death’s ingenious marketing campaign was appealing to sober punks or tech bros who still wanted to feel cool at a gig while holding a non-alcoholic tallboy. The brand quickly won over notable now-sober celebrities like Steve-O, who frequently uses the water on his podcast “Steve-O’s Wild Ride,” and helps the company’s mission to make drinking water cool.
It helps that Cessario is a former creative director for Netflix who knows the power of a good celebrity ad campaign. Last October Cessario recruited Chace Crawford to reprise his character of The Deep (from Amazon’s hit show “The Boys”) to become the company’s “chief sustainability officer”.
Liquid Death’s also recruited comedian Bert Kreischer, adult film actress Cheri DeVille and rapper Wiz Khalifa to do promos. Two years ago, Liquid Death surprised the advertising world by turning negative reviews into a heavy metal album for sale.
The metal album “Greatest Hates” was an attempt to turn bad publicity into sales, and it mostly worked. The album wasn’t a chart-topper, but it certainly got people talking about the product on social media, even the haters. They later doubled down with “Greatest Hates: Vol. 2” the same year, featuring more angry reviews. And a month ago, the brand signed a “pro waterboy” for $100,000 in an act that further solidified their tendency toward irreverent marketing campaigns.
One could even argue that the water brand’s marketing strategy has been so effective, it’s kept most consumers from asking thornier questions about Liquid Death’s business. The company’s calling card is “death to plastic,” but aluminum isn’t exactly sustainable, either.
Still, it remains to be seen if Liquid Death can take over the beverage industry. It is, after all, just one company competing against giants like Nestle, which owns a portion of the upscale water market with holdings in Perrier and San Pellegrino. Not to mention Coca-Cola, whose portfolio includes Dasani, Smart Water, and Topo Chico brands. But none of these brands have the “cool factor” Liquid Death is going for, so maybe its bombastic marketing will give these legacy brands a run for their money.
"After just six months in the market, our flavored sparkling waters are outselling Aha, Bubly, Poland Spring and San Pellegrino in stores," Cessario claimed. "We are the No. 1 dollar contributor to the water category growth over the past year in Whole Foods and are the fastest growing still water brand in Walmart over the last year."
From Your Site Articles
- Liquid Death Raises $23 Million in Series B Funding - dot.LA ›
- How These Beverage Brands Are Shifting Drinking Culture - dot.LA ›
Related Articles Around the Web
Read moreShow less
Samson Amore
Samson Amore is a reporter for dot.LA. He holds a degree in journalism from Emerson College. Send tips or pitches to samsonamore@dot.la and find him on Twitter @Samsonamore.
https://twitter.com/samsonamore
samsonamore@dot.la
In the busy world of entrepreneurship, staying productive is crucial for achieving success. With numerous tasks (and social media) competing for your attention, it can be tough to stay organized and focused. Fortunately, several proven productivity hacks can help you manage your time effectively and prioritize your workload. These strategies are backed by research and have been used by successful individuals in various fields. In this listicle, we’ll explore five effective productivity hacks, including the Pomodoro Technique and the Eisenhower Matrix, that can improve your work habits and enhance your entrepreneurial efforts.
The Pomodoro Technique
Developed by Francesco Cirillo in the late 1980s, the Pomodoro Technique uses a kitchen timer to break work into intervals, typically 25 minutes long, followed by short breaks. Cirillo named the method after the tomato-shaped timer he used as a university student, and it has since been adopted by many professionals and students to enhance focus and productivity. To implement it:
- Choose a task to work on
- Set a timer for 25 minutes
- Work on the task until the timer rings
- Take a short 5-minute break
- After 4 pomodoros, take a longer 15-30 minute break
The Eisenhower Matrix
Named after former U.S. President Dwight D. Eisenhower, this prioritization method helps individuals categorize tasks based on urgency and importance. Eisenhower famously stated, "What is important is seldom urgent and what is urgent is seldom important," which underscores the value of distinguishing between these two aspects to effectively manage time and tasks. This prioritization method helps you focus on what's truly important by categorizing tasks into four quadrants based on urgency and importance:
- Urgent and important: Do immediately
- Important but not urgent: Schedule for later
- Urgent but not important: Delegate if possible
- Neither urgent nor important: Eliminate
Time Blocking
Time blocking is a technique used by many successful entrepreneurs, including Jack Dorsey and Cal Newport, to allocate specific periods for different types of work. By scheduling tasks in advance, individuals can minimize distractions and ensure they dedicate focused time to their most important responsibilities. This creates structure and helps minimize context switching. For example:
- 9-11 AM: Deep work on top priority project
- 11 AM-12 PM: Respond to emails/messages
- 1-3 PM: Meetings
- 3-5 PM: Administrative tasks
The Two-Minute Rule
Popularized by productivity expert David Allen in his book "Getting Things Done," the Two-Minute Rule states that if a task can be completed in two minutes or less, you should do it immediately. This approach helps prevent small tasks from piling up and cluttering your to-do list, allowing for a more organized workflow.
Batching Similar Tasks
Batching tasks is a technique used by many high-performing individuals, including writer Tim Ferriss, who advocates for grouping similar activities to minimize context switching. By focusing on one type of task at a time, such as responding to emails or making phone calls, you can increase efficiency and reduce mental fatigue.
Read moreShow less
RELATEDTRENDING
LA TECH JOBS