HP Tech@Work
Today's trends for tomorrow's business
The Risks of Machine Learning Bias

The Risks of Machine Learning Bias

Way more dangerous than killer robots

This is part three of our machine learning series. You can read part one on machine and deep learning 101 and read part two on biometrics for personal security.
With traditional computing, programmers create an outline for a machine to follow, step by step. With machine learning, computers find their own solutions after crunching vast amounts of data. So how could a computer end up amplifying human prejudices? It’s just data, right?
Let’s step back for a minute and look at an example of machine learning gone wrong. On March 23, 2016 Microsoft released a social chatbot named “Tay” on Twitter. Touted as an experiment in “conversational understanding,” Tay was designed to engage in “casual and playful conversation” with other Twitter users.
The experiment didn’t turn out to be very casual or playful at all. Soon after Tay launched, Twitter trolls started tweeting racist, misogynistic and hateful comments at it; it wasn’t long before Tay began tweeting inflammatory responses right back. Sixteen hours and an astounding 96,000 tweets later, Microsoft pulled the plug on Tay and apologized for the bot’s offensive language.
The Tay debacle was just one of many times that machine-learning software has developed biases based on its given data set. Tay’s flameout didn’t have a long-term impact on society, but there are serious situations where machine-learning bias can have real-world consequences:
  • A ProPublica investigation uncovered that machine learning software designed to spot recidivism in former criminal offenders was twice as likely to flag black men and women as future high-risk criminals; only 20 percent of them actually went on to commit new crimes.
  • Human resources departments hope to use AI to remove human bias when seeking diverse job candidates. However, automatic screening software can latch onto subjective criteria that lead to bias anyway. These screenings backfire when the algorithm decides to seek out, say, European-sounding male names because previous hires had similar names.
  • Some overseas credit-rating companies are using unethical machine learning tactics to analyze the social networks of loan applicants; being connected to someone with good credit is a point in their favor, while connections to friends with poor credit counts against them. This can lead to denied loans, increased car insurance rates and worse.
So if data is biased because the people who compile it are naturally biased, what can we do to curb that bias?

Pledge to do no harm

Cathy O’Neil, a mathematician and author of Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy, proposes that all data scientists should take their own version of the Hippocratic Oath and focus their efforts on ways to regulate machine learning math models to make them fairer.

Launch broad ethics oversights

As machine learning advances, its effects will be felt throughout society - so it makes sense that its regulation shouldn’t be left entirely to engineers and tech companies. Watchdog groups such as the Ethics and Governance of Artificial Intelligence Fund gets input from social scientists, ethicists, philosophers, faith leaders, economists, lawyers and more to make sure a wide range of perspectives are incorporated into their ethical guidelines.

Diversify the AI industry

Representation matters, especially when you’re trying to remove unconscious biases baked into data sets. Organizations such as Women in Machine Learning and Black in AI mentor and encourage the advancement of people underrepresented in the field. And just this month the African Institute for Mathematical Sciences announced a new machine learning master’s degree program, the first of its kind in Africa, which Facebook and Google are helping fund.

Get started

Explore machine learning on machines designed for it - HP Z4 and Z8 workstations - and visit HP®’s machine learning portal for developers to install a high performance, fully tested machine learning environment onto your HP Z workstation.

Disclosure: Our site may get a share of revenue from the sale of the products featured on this page.

More about these products
Disclaimer

Prices, specifications, availability and terms of offers may change without notice. Price protection, price matching or price guarantees do not apply to Intra-day, Daily Deals or limited-time promotions. Quantity limits may apply to orders, including orders for discounted and promotional items. Despite our best efforts, a small number of items may contain pricing, typography, or photography errors. Correct prices and promotions are validated at the time your order is placed. These terms apply only to products sold by HP.com; reseller offers may vary. Items sold by HP.com are not for immediate resale. Orders that do not comply with HP.com terms, conditions, and limitations may be cancelled. Contract and volume customers not eligible.

HP’s MSRP is subject to discount. HP’s MSRP price is shown as either a stand-alone price or as a strike-through price with a discounted or promotional price also listed. Discounted or promotional pricing is indicated by the presence of an additional higher MSRP strike-through price

The following applies to HP systems with Intel 6th Gen and other future-generation processors on systems shipping with Windows 7, Windows 8, Windows 8.1 or Windows 10 Pro systems downgraded to Windows 7 Professional, Windows 8 Pro, or Windows 8.1: This version of Windows running with the processor or chipsets used in this system has limited support from Microsoft. For more information about Microsoft’s support, please see Microsoft’s Support Lifecycle FAQ at https://support.microsoft.com/lifecycle

Ultrabook, Celeron, Celeron Inside, Core Inside, Intel, Intel Logo, Intel Atom, Intel Atom Inside, Intel Core, Intel Inside, Intel Inside Logo, Intel vPro, Itanium, Itanium Inside, Pentium, Pentium Inside, vPro Inside, Xeon, Xeon Phi, Xeon Inside, and Intel Optane are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries.

In-home warranty is available only on select customizable HP desktop PCs. Need for in-home service is determined by HP support representative. Customer may be required to run system self-test programs or correct reported faults by following advice given over phone. On-site services provided only if issue can't be corrected remotely. Service not available holidays and weekends.

HP will transfer your name and address information, IP address, products ordered and associated costs and other personal information related to processing your application to Bill Me Later®. Bill Me Later will use that data under its privacy policy.

Microsoft Windows 10: Not all features are available in all editions or versions of Windows 10. Systems may require upgraded and/or separately purchased hardware, drivers, software or BIOS update to take full advantage of Windows 10 functionality. Windows 10 is automatically updated, which is always enabled. ISP fees may apply and additional requirements may apply over time for updates. See http://www.microsoft.com.

“Best All In One Printer” and “the easiest printer you’ve ever had to set up” from Wirecutter. ©2020 The Wirecutter, Inc.. All rights reserved. Used under license. https://www.nytimes.com/wirecutter/reviews/best-all-in-one-printer/

Get Marvel’s Avengers when you purchase HP gaming PCs with qualifying 9th gen or 10th gen Intel® Core™ i5, i7 and i9 processors. Redemption code will be sent out by email within 60 days of purchase. Limited quantities and while supply lasts. Offer valid thru 12/31/2020 only while supplies last. We reserve the right to replace titles in the offer for ones of equal or greater value. Certain titles may not be available to all consumers because of age restrictions. The Offer may be changed, cancelled, or suspended at any time, for any reason, without notice, at Intel’s reasonable discretion if its fairness or integrity affected whether due to human or technical error. The Offer sponsor is Intel Corporation, 2200 Mission College Blvd., Santa Clara, CA 95054, USA. To participate you must create an Intel Digital Hub Account, purchase a qualifying product during the redemption period, enter a valid Master Key, and respond to a brief survey. Information you submit is collected, stored, processed, and used on servers in the USA. For more information on offer details, eligibility, restrictions, and our privacy policy, visit https://softwareoffer.intel.com/offer/20Q3-19/terms.

© 2020 MARVEL. © Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries in the U.S. and/or other countries. Other names and brands may be claimed as the property of others.

The personal information you provide will be used according to the HP Privacy Statement (https://www8.hp.com/us/en/privacy/ww-privacy.html)