Connect with us

Ai

Oh dear… AI models used to flag hate speech online are, er, racist against black people

Published

on

The internet is filled with trolls spewing hate speech, but machine learning algorithms can’t help us clean up the mess.

A paper from computer scientists from the University of Washington, Carnegie Mellon University, and the Allen Institute for Artificial Intelligence, found that machines were more likely to flag tweets from black people than white people as offensive. It all boils down to the subtle differences in language. African-American English (AAE), often spoken in urban communities, is peppered with racial slang and profanities.

But even if they contain what appear to be offensive words, the message itself often isn’t abusive. For example, the tweet “I saw him yesterday” is scored as 6 per cent toxic, but it suddenly skyrockets to 95 per cent for the comment “I saw his ass yesterday”. The word ass may be crude, but when used in that context it’s not aggressive at all.

racial_bias_hate_speech_detection

An example of how African-American English (AAE) is mistakenly classified as offensive compared to standard American English. Image credit: Sap et al.

“I wasn’t aware of the exact level of bias in Perspective API–the tool used to detect online hate speech–when searching for toxic language, but I expected to see some level of bias from previous work that examined how easily algorithms like AI chatter bots learn negative cultural stereotypes and associations,” said Saadia Gabriel, co-author of the paper and a PhD student at the University of Washington.

“Still, it’s always surprising and a little alarming to see how well these algorithms pick up on toxic patterns pertaining to race and gender when presented with large corpora of unfiltered data from the web.”

The researchers fed a total of 124,779 tweets collected from two datasets that were classified as toxic according to Perspective API. Originally developed by Google and Jigsaw, an incubator company currently operating under Alphabet, the machine learning software is used by Twitter to flag any abusive comments.

The tool mistakenly classified 46 per cent of non-offensive tweets crafted in the style of African American English (AAE) as inflammatory, compared to just nine per cent of tweets written in standard American English.

“I think we have to be really careful about what technologies we implement in general, whether it’s a platform where people can post whatever they want, or whether is an algorithm that detects certain types of (potentially harmful) content. Platforms are under increasing pressure to delete harmful content, but currently these deletions are backfiring against minorities,” Maarten Sap, first author of the paper and a PhD student at the University of Washington, told The Register.

When humans were employed via the Amazon Mechanical Turk service to look at 1,351 tweets from the same dataset and asked to judge if the comment was either offensive to them or could be seen as offensive to anyone.

Just over half – about 55 per cent – were classified as “could be offensive to anyone”. That figure dropped to 44 per cent, however, when they were asked to consider the user’s race and their use of AAE.

shock_computer

Q. If machine learning is so smart, how come AI models are such racist, sexist homophobes? A. Humans really suck

READ MORE

“Our work serves as a reminder that hate speech and toxic language is highly subjective and contextual,” said Sap.

“We have to think about dialect, slang and in-group versus out-group, and we have to consider that slurs spoken by the out-group might actually be reclaimed language when spoken by the in-group.”

The study provides yet another reminder that AI models don’t understand the world enough to have common sense. Tools like Perspective API often fail when faced with subtle nuances in human language or even incorrect spellings.

Similar models employed by other social media platforms like Facebook to detect things like violence or pornography often don’t work for the same reason. And this is why these companies can’t rely on machines alone, and have to hire teams of human contractors to moderate questionable content.

Sap believes that removing the humans from content moderation isn’t the way to go.

“We managed to reduce some of the bias by making workers more aware of the existence of African American English, and reminding them that certain seemingly obscene words could be harmless depending on who speaks them. Knowing how flawed humans are at this task, especially given the working conditions that some companies put their content moderators in, I certainly don’t think humans are flawless in this capacity. However, I don’t think removing them from the equation is necessarily the way to go either. I think a good collaborative human+AI setting is likely the best option, but only time will tell.” ®

Sponsored: Detecting cyber attacks as a small to medium business

Source: https://go.theregister.co.uk/feed/www.theregister.co.uk/2019/10/11/ai_black_people/

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published.

Ai

Two Sigma Ventures raises $288M, complementing its $60B hedge fund parent

Published

on

By

Eight years ago, Two Sigma Investments began an experiment in early-stage investing.

The hedge fund, focused on data-driven quantitative investing, was well on its way to amassing the $60 billion in assets under management that it currently holds, but wanted more exposure to early-stage technology companies, so it created a venture capital arm, Two Sigma Ventures.

At the time of the firm’s launch it made a series of investments, totaling about $70 million, exclusively with internal capital. The second fund was a $150 million vehicle that was backed primarily by the hedge fund, but included a few external limited partners.

Now, eight years and several investments later, the firm has raised $288 million in new funding from outside investors and is pushing to prove out its model, which leverages its parent company’s network of 1,700 data scientists, engineers and industry experts to support development inside its portfolio.

The world is becoming awash in data and there’s continuing advances in the science of computing,” says Two Sigma Ventures co-founder Colin Beirne. “We thought eight years ago when when started, that more and more companies of the future would be tapping into those trends.”

Beirne describes the firm’s investment thesis as being centered on backing data-driven companies across any sector — from consumer technology companies like the social networking monitoring application, Bark, or the high-performance, high-end sports wearable company, Whoop.

Alongside Beirne, Two Sigma Ventures is led by three other partners: Dan Abelon, who co-founded SpeedDate and sold it to IAC; Lindsey Gray, who launched and led NYU’s Entrepreneurial Institute; and Villi Iltchev, a former general partner at August Capital.

Recent investments in the firm’s portfolio include Firedome, an endpoint security company; NewtonX, which provides a database of experts; Radar, a location-based data analysis company; and Terray Therapeutics, which uses machine learning for drug discovery.

Other companies in the firm’s portfolio are farther afield. These include the New York-based Amper Music, which uses machine learning to make new music; and Zymergen, which uses machine learning and big data to identify genetic variations useful in pharmaceutical and industrial manufacturing.

Currently, the firm’s portfolio is divided between enterprise investments, consumer-facing deals and healthcare-focused technologies. The biggest bucket is enterprise software companies, which Beirne estimates represents about 65% of the portfolio. He expects the firm to become more active in healthcare investments going forward.

“We really think that the intersection of data and biology is going to change how healthcare is delivered,” Beirne says. “That looks dramatically different a decade from now.”

To seed the market for investments, the firm’s partners have also backed the Allen Institute’s investment fund for artificial intelligence startups.

Together with Sequoia, KPCB and Madrona, Two Sigma recently invested in a $10 million financing to seed companies that are working with AI. “This is a strategic investment from partner capital,” says Beirne.

Typically startups can expect Two Sigma to invest between $5 million and $10 million with its initial commitment. The firm will commit up to roughly $15 million in its portfolio companies over time.

Read more: https://techcrunch.com/2020/01/22/two-sigma-ventures-raises-288-million-complementing-its-60-billion-hedge-fund-parent/

Continue Reading

Ai

German football league Bundesliga teams with AWS to improve fan experience

Published

on

By

Germany’s top soccer (football) league, Bundesliga, announced today it is partnering with AWS to use artificial intelligence to enhance the fan experience during games.

Andreas Heyden, executive vice president for digital sports at the Deutsche Fußball Liga, the entity that runs Bundesliga, says that this could take many forms, depending on whether the fan is watching a broadcast of the game or interacting online.

“We try to use technology in a way to excite a fan more, to engage a fan more, to really take the fan experience to the next level, to show relevant stats at the relevant time through broadcasting, in apps and on the web to personalize the customer experience,” Heyden said.

This could involve delivering personalized content. “In times like this when attention spans are shrinking, when a user opens up the app the first message should be the most relevant message in that context in that time for the specific user,” he said.

It also can help provide advanced statistics to fans in real time, even going so far as to predict the probability of a goal being scored at any particular moment in a game that would have an impact on your team. Heyden thinks of it as telling a story with numbers, rather than reporting what happened after the fact.

“We want to, with the help of technology, tell stories that could not have been told without the technology. There’s no chance that a reporter could come up with a number of what the probability of a shot [scoring in a given moment]. AWS can,” he said.

Werner Vogels, CTO at Amazon, says this about using machine learning and other technologies on the AWS platform to add to the experience of watching the game, which should help attract younger fans, regardless of the sport. “All of these kind of augmented customer fan experiences are crucial in engaging a whole new generation of fans,” Vogels told TechCrunch.

He adds that this kind of experience simply wasn’t possible until recently because the technology didn’t exist. “These things were impossible five or 10 years ago, mostly because now with all the machine learning software, as well as how the [pace of technology] has accelerated at such a [rate] at AWS, we’re now able to do these things in real time for sports fans.”

Bundesliga is not just any football league. It is the second biggest in the world in terms of revenue, and boasts the highest stadium attendance of all football teams worldwide. Today’s announcement is an extension of an ongoing relationship between DFL and AWS, which started in 2015 when Heyden helped move the league’s operations to the cloud on AWS.

Heyden says that it’s not a coincidence he ended up using AWS instead of another cloud company. He has known Vogels (who also happens to be a huge soccer fan) for many years, and has been using AWS for more than a decade, even well before he joined the DFL. Today’s announcement is an extension of that long-term relationship.

Read more: https://techcrunch.com/2020/01/24/german-football-league-bundesliga-teams-with-aws-to-improve-fan-experience/

Continue Reading

Ai

Delta Air Lines startup partnerships are fueling innovation

Published

on

By

For the first time, this year Delta Air Lines had a large presence at CES. The carrier used much of its space to highlight the “parallel reality” screens developed by Misapplied Sciences and Sarcos Robotics, which brought its latest Guardian exoskeleton. At the show, I sat down with COO Gil West, an industry veteran with years of experience at a number of airlines and airplane manufacturers, to talk about how the company works with these startups.

Like all large companies, Delta has gone through a bit of a digital transformation in recent years by rebuilding a lot of the technical infrastructure that powers its internal and external services (though like all airlines, it also still has plenty of legacy tech that is hard to replace). This work enabled the company to move faster, rethink a lot of its processes and heightened the reality that a lot of this innovation has to come from outside the company.

“If you think about where we are as a world right now, it’s a Renaissance period for transportation,” West said. “Now, fortunately, we’re right in the middle of it, but if you think about the different modes of transportation and autonomous and electrification — and the technologies like AI and ML — everything is converging. There’s truly, I think, a transportation revolution — and we’ll play in it.

Read more: https://techcrunch.com/2020/01/14/delta-air-lines-startup-partnerships-are-fueling-innovation/

Continue Reading

Trending