Kendra Gaunt (she/her or they/them pronouns) is a data and AI product owner at The Trevor Project, the world’s largest suicide prevention and crisis intervention organization for LGBTQ youth. A 2019 Google AI Impact Grantee, the organization is implementing new AI applications to scale its impact and save more young LGBTQ lives.
By now, most of us in tech know that the inherent bias we possess as humans creates an inherent bias in AI applications — applications that have become so sophisticated they’re able to shape the nature of our everyday lives and even influence our decision-making.
The more prevalent and powerful AI systems become, the sooner the industry must address questions like: What can we do to move away from using AI/ML models that demonstrate unfair bias?
How can we apply an intersectional framework to build AI for all people, knowing that different individuals are affected by and interact with AI in different ways based on the converging identities they hold?
Start with identifying the variety of voices that will interact with your model. Intersectionality: What it means and why it matters
Before tackling the tough questions, it’s important to take a step back and define “intersectionality.” A term defined by Kimberlé Crenshaw, it’s a framework that empowers us to consider how someone’s distinct identities come together and shape the ways in which they experience and are perceived in the world.
This includes the resulting biases and privileges that are associated with each distinct identity. Many of us may hold more than one marginalized identity and, as a result, we’re familiar with the compounding effect that occurs when these identities are layered on top of one another.
At The Trevor Project, the world’s largest suicide prevention and crisis intervention organization for LGBTQ youth, our chief mission is to provide support to each and every LGBTQ young person who needs it, and we know that those who are transgender and nonbinary and/or Black, Indigenous, and people of color face unique stressors and challenges.
So, when our tech team set out to develop AI to serve and exist within this diverse community — namely to better assess suicide risk and deliver a consistently high quality of care — we had to be conscious of avoiding outcomes that would reinforce existing barriers to mental health resources like a lack of cultural competency or unfair biases like assuming someone’s gender based on the contact information presented.
Though our organization serves a particularly diverse population, underlying biases can exist in any context and negatively impact any group of people. As a result, all tech teams can and should aspire to build fair, intersectional AI models, because intersectionality is the key to fostering inclusive communities and building tools that serve people from all backgrounds more effectively.
Doing so starts with identifying the variety of voices that will interact with your model, in addition to the groups for which these various identities overlap. Defining the opportunity you’re solving is the first step because once you understand who is impacted by the problem, you can identify a solution. Next, map the end-to-end experience journey to learn the points where these people interact with the model. From there, there are strategies every organization, startup and enterprise can apply to weave intersectionality into every phase of AI development — from training to evaluation to feedback.
Datasets and training
The quality of a model’s output relies on the data on which it’s trained. Datasets can contain inherent bias due to the nature of their collection, measurement and annotation — all of which are rooted in human decision-making. For example, a 2019 study found that a healthcare risk-prediction algorithm demonstrated racial bias because it relied on a faulty dataset for determining need. As a result, eligible Black patients received lower risk scores in comparison to white patients, ultimately making them less likely to be selected for high-risk care management.
Fair systems are built by training a model on datasets that reflect the people who will be interacting with the model. It also means recognizing where there are gaps in your data for people who may be underserved. However, there’s a larger conversation to be had about the overall lack of data representing marginalized people — it’s a systemic problem that must be addressed as such, because sparsity of data can obscure both whether systems are fair and whether the needs of underrepresented groups are being met.
To start analyzing this for your organization, consider the size and source of your data to identify what biases, skews or mistakes are built-in and how the data can be improved going forward.
The problem of bias in datasets can also be addressed by amplifying or boosting specific intersectional data inputs, as your organization defines it. Doing this early on will inform your model’s training formula and help your system stay as objective as possible — otherwise, your training formula may be unintentionally optimized to produce irrelevant results.
At The Trevor Project, we may need to amplify signals from demographics that we know disproportionately find it hard to access mental health services, or for demographics that have small sample sizes of data compared to other groups. Without this crucial step, our model could produce outcomes irrelevant to our users.
Model evaluation is an ongoing process that helps organizations respond to ever-changing environments. Evaluating fairness began with looking at a single dimension — like race or gender or ethnicity. The next step for the tech industry is figuring out how to best compare intersectional groupings to evaluate fairness across all identities.
To measure fairness, try defining intersectional groups that could be at a disadvantage and the ones that may have an advantage, and then examine whether certain metrics (for example, false-negative rates) vary among them. What do these inconsistencies tell you? How else can you further examine which groups are underrepresented in a system and why? These are the kinds of questions to ask at this phase of development.
Developing and monitoring a model based on the demographics it serves from the start is the best way for organizations to achieve fairness and alleviate unfair bias. Based on the evaluation outcome, a next step might be to purposefully overserve statistically underrepresented groups to facilitate training a model that minimizes unfair bias. Since algorithms can lack impartiality due to societal conditions, designing for fairness from the outset helps ensure equal treatment of all groups of individuals.
Feedback and collaboration
Teams should also have a diverse group of people involved in developing and reviewing AI products — people who are diverse not only in identities, but also in skillset, exposure to the product, years of experience and more. Consult stakeholders and those who are impacted by the system for identifying problems and biases.
Lean on engineers when brainstorming solutions. For defining intersectional groupings, at The Trevor Project, we worked across the teams closest to our crisis-intervention programs and the people using them — like Research, Crisis Services and Technology. And reach back out to stakeholders and people interacting with the system to collect feedback upon launch.
Ultimately, there isn’t a “one-size-fits-all” approach to building intersectional AI. At The Trevor Project, our team has outlined a methodology based on what we do, what we know today and the specific communities we serve. This is not a static approach and we remain open to evolving as we learn more. While other organizations may take a different approach to build intersectional AI, we all have a moral responsibility to construct fairer AI systems, because AI has the power to highlight — and worse, magnify — the unfair biases that exist in society.
Depending on the use case and community in which an AI system exists, the magnification of certain biases can result in detrimental outcomes for groups of people who may already face marginalization. At the same time, AI also has the ability to improve quality of life for all people when developed through an intersectional framework. At The Trevor Project, we strongly encourage tech teams, domain experts and decision-makers to think deeply about codifying a set of guiding principles to initiate industry-wide change — and to ensure future AI models reflect the communities they serve.
Resistant AI scores $16.6M for its anti-fraud fintech tools – TechCrunch
Resistant AI, which uses artificial intelligence to help financial services companies combat fraud and financial crime — selling tools to protect credit risk scoring models, payment systems, customer onboarding and more — has closed $16.6 million in Series A funding. GV (formerly Google Ventures) led the round, with participation from existing investors Index Ventures (led […]
Resistant AI, which uses artificial intelligence to help financial services companies combat fraud and financial crime — selling tools to protect credit risk scoring models, payment systems, customer onboarding and more — has closed $16.6 million in Series A funding.
GV (formerly Google Ventures) led the round, with participation from existing investors Index Ventures (led by partner Jan Hammer), Credo Ventures (led by Ondrej Bartos and Vladislav Jez) and Seedcamp, plus several unnamed angel investors specializing in financial technology and security.
The 2019-founded, Prague-based startup says the funding will be used to meet rising demand from global financial institutions, including by building out its product, engineering, and sales operations teams beyond its existing footprint — which also includes offices in London and New York.
The startup tells TechCrunch it has 30 customers signed up at this stage to use its dedicated anti-fraud security products — which include machine learning detection of fraudulent documents and AI for spotting problematic patterns of transactions.
Collectively its customers, which include banks, insurance companies and fintechs — it can’t name the biggest but names the likes of KBC, Payoneer, Habito and Twisto — are processing tens of millions of transactions per month, it also said, adding that in its home market of the Czech Republic it’s now working with banks that have a combined 50% marketshare.
To give a taster of the problem it’s tackling, the startup says that assessment of customer data it’s acquired and reviewed indicates: 17% of bank statement that are used for lending applications, ‘Know Your Customer’ regulations and other purposes are tampered with; 11% of UK payslips submitted as part of digital loan applications are altered or forged; 15% of company registration certificates submitted worldwide when opening a bank account are fakes; and 9% of utility bills submitted as a proof of address worldwide are forged.
“Our mission is to create an intelligent shield for autonomous financial systems, to protect them against these ever-evolving, ever-smarter attacks,” adds CEO Martin Rehak in a statement. “That’s the only way we can avoid epidemic fraud, mountains of manual reviews and four-factor authentication on every single online service.”
In another supporting statement, Tom Hulme, general partner at GV, said: “Resistant AI’s founding team has unique expertise in applying AI and machine learning to detect complex and evasive behavior. Early customer traction demonstrates an ability to uncover unknown threats, and reliably categorize and reduce false alerts with transparent, explainable and verifiable detection models.”
Facebook reportedly plans to change its name to focus on the metaverse – TechCrunch
Facebook is planning to rebrand the company with a new name to focus on building the metaverse, according to a report by The Verge. CEO Mark Zuckerberg will unveil its new name at the annual Connect conference on October 28, but it could announce the new name earlier, as reported by The Verge. Facebook, which […]
Facebook is planning to rebrand the company with a new name to focus on building the metaverse, according to a report by The Verge.
CEO Mark Zuckerberg will unveil its new name at the annual Connect conference on October 28, but it could announce the new name earlier, as reported by The Verge.
Facebook, which has the ambition to be known for more than social media, announced Sunday that it plans to recruit 10,000 jobs in Europe for the next five years to help build the metaverse the company sees as a key component of its future.
The company also announced a month ago that Andre Bosworth, the head of AR and VR, will be promoted to chief technology officer. Facebook already has more than 10,000 employees who build consumer hardware like AR glasses that Zuckerberg believes will be as ubiquitous as smartphones.
In July, Zuckerberg said that Facebook’s future lies in the virtual metaverse, in which users will live, work and play inside.
The rebranding comes at a time when Facebook is facing criticism over a range of scandals, including a series of internal documents leaked by a whistleblower, Frances Haugen, who testified before the Senate Committee on Commerce, Science, and Transportation. Facebook is still under antitrust scrutiny by the U.S. government.
“We don’t comment on rumor or speculation,” a Facebook spokesperson said.
Does the NFT craze actually matter? – TechCrunch
Hello friends, and welcome back to Week in Review! Last week, we talked about Apple’s subscription addiction. This week, I’m diving deep into whether there’s actually any meaning to pull out of the NFT mania of 2021. If you’re reading this on the TechCrunch site, you can get this in your inbox from the newsletter […]
Hello friends, and welcome back to Week in Review!
Last week, we talked about Apple’s subscription addiction. This week, I’m diving deep into whether there’s actually any meaning to pull out of the NFT mania of 2021.
the big thing
The NFT market is still defying reason, but then again that’s kind of its thing. But one thing I’m especially unsure about lately as I see JPGs continue to sell for millions of dollars is… does any of this actually matter?
I’ve spent a lot of time over the last year grappling with the NFT market, at times I’ve lost sleep over it. As a reporter frequently covering this market, I don’t own or trade the little images myself, but that hasn’t stopped me from obsessing over the fluctuations in their prices and scouring Discords trying to follow the trends. I’ve tuned into countless Twitter Spaces and lurked subreddits trying to understand it all. I’ve also done my best to keep most of that out of this newsletter — it’s a weird niche interest that’s especially niche at the moment — but as Bitcoin flirts with a new all-time-high and the NFT mania persists, just consider this a timely update.
So, in the past month, investors have continued dropping billions upon billions of dollars on NFTs. OpenSea has seen more than $3 billion in transaction volume in the past 30 days, and that number is actually way down quite a bit from August, showcasing just how much off-peak money continues to flow into NFTs.
All of that money has gone to some colorful places. One of the bigger success stories of the past month has been the platform CrypToadz which investors dumped $100 million into. They look like this. In the past couple weeks, a brand new project called MekaVerse saw $130 million in transaction volume. They’re a bit prettier, but would you spend more than $8,000 on one? The platform Cryptoslam (where I pulled most of the data I reference here) is tracking 163 platforms which did more than $1 million in volume in the past 30 days, a number which doesn’t even account for individual artists selling their work on platforms like OpenSea.
Now, there are two incredibly different segments of NFT communities out there, larger-scale NFT projects like Axie Infinity and NBA Top Shot with tens and hundreds of thousands of users and smaller-scale NFT projects like CryptoPunks and Art Blocks with just a few hundred or thousand owners. Larger-scale projects can represent more traditional gaming titles with more complex in-game economies while smaller-scale projects simply look more like fine art markets teamed with exclusive social clubs. Some smaller-scale projects have the ambition to eventually become larger-scale ones, but many have capped the number of NFTs in their projects and are designed to be exclusive.
In the past 30 days, Axie Infinity did more than $500 million in sales spread across nearly 2 million transactions and over 350,000 buyers. On the flip side, CryptoPunks did $200 million in sales during that same time frame across 484 transactions and 309 buyers.
Generally, when I’m talking about some of these big sales from smaller-scale projects with friends of mine, the first thing they mention is how this is probably all just money laundering. While I’d certainly imagine some of that is happening, that’s ultimately a much more boring explanation than my best guess of what’s really going on, which is that a group of several thousand investors have separately rationalized irrational investing. They just happen to have chosen to do so through buying pixel art and drawings of animals.
While some investors might suggest that a handful of the earliest NFTs hold intrinsic value as historic objects, there are plenty of brand new NFT projects earning ten-million dollar valuations on day one with low amounts of effort and imagination.
It’s seemingly the result of momentum from awe-struck retail investors entering a market filled with massive amounts of wealth being generated and re-invested by Ethereum millionaires who can massively overpay for deals while pushing the implied value of the objects, the projects, the entire NFT market and the price of Ethereum up concurrently. Most of these investors are also people who have held onto Ethereum through its waves and have grown fundamentally averse to cashing out, meaning they’re less likely to sell the NFTs they buy unless they’re just trying to buy another more expensive NFT or have been made an offer too good to refuse. As a result, many high-value smaller-scale projects stay liquid on the low-end while fewer sales of the rarer items underpin the massive valuations of the projects and those occasional big buys keep pushing prices higher.
All of this babbling of mine is to say, what’s happening here is strange. It’s also an incredibly large amount of noise mostly coming from a few thousand buyers.
But when most investors talk about mainstream adoption and future use cases, they’re looking at the creation of more larger-scale projects like Axie and Top Shot which embody many of the technical bells and whistles of crypto economics in more user-friendly packages that can reach the mainstream. NFTs as a concept for driving more complex virtual economies is, indeed, really fascinating, but I don’t think there are as many takeaways to draw from billions of dollars flowing into digital art and these smaller-scale projects like CrypToadz as many crypto investors and venture capitalists are trying to convince themselves.
Only three NFT platforms out there had more than 10,000 active unique buyers in their community in the past 30 days, and while the successes of platforms like Axie Infinity are definitely worth dissecting, it also seems clear we’re in the midst of a speculative frenzy and it’s not a very easy time to draw sober conclusions about what all this madness means for the future of the web.
Here are a few stories this week that I think you should take a closer look at:
Apple probably won’t be supporting alternate App Store payments anytime soon
Apple did their best to convince the press and public that the court’s decision in its legal fight with Epic Games was an outright win for Apple, but over the weekend they quietly announced that weeks later they’re appealing the decision and asking the courts to put the ordered changes to allow alternative payments inside iOS apps on hold.
Apple put on a cool demeanor after this ruling, but it’s apparent that there are billions on the line for Apple if this order stands. Therefore delaying its rollout means billions of dollars that aren’t going to other payment providers or staying in developer coffers. Epic had already appealed the decision as well, hoping to try for a more favorable ruling, but it’s clear that anyone hoping for a speedy resolution will be disappointed — as is often the case in corporate law.
Nintendo reshapes its SaaS ambitions
Nintendo has been and probably always will be a bit of an odd big company. They’ve been resistant to new trends in gaming and when they embrace them, they don’t necessarily do a great job capitalizing on them, and yet their mountain of beloved IP allows them chance after chance to get things right. This week, they announced more details on their new annual membership called Nintendo Switch Online+ which, for $50 per year will give gamers a deeper array of content. That’s a good deal more than the standard $20 per year for the regular Nintendo Switch Online subscription, but beyond expanded virtual console support for an unannounced array of N64 games, it’s not clear what exactly the sell is for consumers.
Interestingly, they’re launching the service with free access to a major update for Animal Crossing: New Horizons. It’s a play that only works when you’re Nintendo and the penetration of your first-party titles is so incredibly high among device-owners (and especially likely subscribers). Nintendo has sold more than 3.4 million copies of the new Animal Crossing title globally.
Microsoft pulls LinkedIn from China
It’s been a particularly turbulent time for tech companies across China as government regulators crack down and the outlook clouds for big platforms there. This week, Microsoft announced that it’s pulling LinkedIn out of China, detailing that LinkedIn was now “facing a significantly more challenging operating environment and greater compliance requirements in China.” LinkedIn didn’t have a huge presence in China so this won’t make major waves, but as other American tech giants are forced to make major adjustments to their China strategy, this marks yet another datapoint in the cooling of relations between China and the West.
The LinkedIn’s of the world don’t hold much sway in China, the most curious bit of this is how this regulatory upswing eventually affects Apple which does hold plenty of influence. While officials probably aren’t keen to jam them up, the past year has shown that China’s regulators have plenty of surprises up their sleeves.
Some of my favorite reads from our newly-renamed TechCrunch+ subscription service this week:
“…Visa and Plaid might have chosen to go their own ways in the end, but the year wasn’t a total loss for the data connectivity startup: Plaid claims its customer count grew 60% in 2020, and company execs say it has had similar growth so far this year….”
Founders should use predictive modeling to fundraise smarter
“More capital is flooding into growth equity at earlier stages, and it’s happening faster than ever before. But even with the rampant enthusiasm for pouring bigger equity checks into startups, founders are now in a unique place in time where they can think differently about how to capitalize their companies….“
How one startup boosted productivity with ‘get s*** done’ day
“…To improve our productivity, we introduced a Getting Shit Done Day (GSDD): Our employees define clear-cut goals and receive specific, usually non-trivial, tasks with little to no communication involved (we encourage our employees to avoid social media on this day, but we are not looking over their shoulder). The goal of GSDD is to increase the amount of time we spend in deep work by minimizing distractions for one day every other week…”
Eco-friendly sneaker maker Allbirds aims for $2 bln valuation in U.S. IPO
Eco-friendly sneaker maker Allbirds Inc said on Monday it aims to be valued at over $2 billion in its New...
Comcast gave me good, precise news. The truth was precisely the opposite
Many companies believe that technology is perfect for customer service communication. Often, though, it just isn't.
Sex differences in COVID-19 outcomes
Credit: Mary Ann Liebert, Inc., publishers In a study of more than 10,600 adult patients hospitalized with COVID-19, women had
Resistant AI scores $16.6M for its anti-fraud fintech tools – TechCrunch
Resistant AI, which uses artificial intelligence to help financial services companies combat fraud and financial crime — selling tools to...
Facebook reportedly plans to change its name to focus on the metaverse – TechCrunch
Facebook is planning to rebrand the company with a new name to focus on building the metaverse, according to a...
UTHSC awarded $1.5 million HRSA grant for sexual assault nurse examiner training
Credit: UTHSC Memphis, Tenn. (June 16, 2021) - The University of Tennessee Health Science Center's College of Nursing has received
Does the NFT craze actually matter? – TechCrunch
Hello friends, and welcome back to Week in Review! Last week, we talked about Apple’s subscription addiction. This week, I’m...
Corporate Company Earnings, Find Earnings Per Share and Earnings History Online
Even computer experts think ending human oversight of AI is a very bad idea
The UK government is thinking of scrapping the right to ask for a human to review decisions made entirely by...
The Briefing: Hailo Lands $136M Series C
Crunchbase News' top picks of the news to stay current in the VC and startup world.
Entrepreneur3 months ago
How Success Happened for Josh Harris, Co-Founder of Apollo Global Management and Co-Founder of Harris Blitzer Sports & Entertainment
Ethicalmarkets11 months ago
Finance for Biodiversity – Aligning Development Finance with Nature’s Needs: Protecting Nature’s Development Dividend
Bioengineer11 months ago
Scientists warn of the social and environmental risks tied to the energy transition
Business insider5 months ago
Prime Minister of Dominica Provides Update on International Airport and Importance of Citizenship by Investment Funds to Public Sector
Ethicalmarkets12 months ago
Third TCFD Status Report Shows Progress & Highlights Need for Greater Climate-Related Disclosures and Transparency
Bioengineer3 months ago
$1 million grant to address cold storage logistics in vaccine delivery
Bioengineer4 months ago
Particles with ‘eyes’ allow a closer look at rotational dynamics
Business insider12 months ago
10 things you need to know before the opening bell | Markets Insider