In debates over the future of artificial intelligence, many experts think of the new systems as coldly logical and objectively rational. But in a new study, researchers have demonstrated how machines can be reflections of us, their creators, in potentially problematic ways.
Common machine learning programs, when trained with ordinary human language available online, can acquire cultural biases embedded in the patterns of wording, the researchers found. These biases range from the morally neutral, like a preference for flowers over insects, to the objectionable views of race and gender.
Identifying and addressing possible bias in machine learning will be critically important as we increasingly turn to computers for processing the natural language humans use to communicate, for instance in doing online text searches, image categorization and automated translations.
“Questions about fairness and bias in machine learning are tremendously important for our society,” said researcher Arvind Narayanan, an assistant professor of computer science and an affiliated faculty member at the Center for Information Technology Policy (CITP) at Princeton University, as well as an affiliate scholar at Stanford Law School’s Center for Internet and Society. “We have a situation where these artificial intelligence systems may be perpetuating historical patterns of bias that we might find socially unacceptable and which we might be trying to move away from.”
The paper, “Semantics derived automatically from language corpora contain human-like biases,” published April 14 in Science. Its lead author is Aylin Caliskan, a postdoctoral research associate and a CITP fellow at Princeton; Joanna Bryson, a reader at University of Bath, and CITP affiliate, is a coauthor.
As a touchstone for documented human biases, the study turned to the Implicit Association Test, used in numerous social psychology studies since its development at the University of Washington in the late 1990s. The test measures response times (in milliseconds) by human subjects asked to pair word concepts displayed on a computer screen. Response times are far shorter, the Implicit Association Test has repeatedly shown, when subjects are asked to pair two concepts they find similar, versus two concepts they find dissimilar.
Take flower types, like “rose” and “daisy,” and insects like “ant” and “moth.” These words can be paired with pleasant concepts, like “caress” and “love,” or unpleasant notions, like “filth” and “ugly.” People more quickly associate the flower words with pleasant concepts, and the insect terms with unpleasant ideas.
The Princeton team devised an experiment with a program where it essentially functioned like a machine learning version of the Implicit Association Test. Called GloVe, and developed by Stanford University researchers, the popular, open-source program is of the sort that a startup machine learning company might use at the heart of its product. The GloVe algorithm can represent the co-occurrence statistics of words in, say, a 10-word window of text. Words that often appear near one another have a stronger association than those words that seldom do.
The Stanford researchers turned GloVe loose on a huge trawl of contents from the World Wide Web, containing 840 billion words. Within this large sample of written human culture, Narayanan and colleagues then examined sets of so-called target words, like “programmer, engineer, scientist” and “nurse, teacher, librarian” alongside two sets of attribute words, such as “man, male” and “woman, female,” looking for evidence of the kinds of biases humans can unwittingly possess.
The Latest on: Bias in artificial intelligence systems
via Google News
The Latest on: Bias in artificial intelligence systems
- RiskSense CEO Invited to Moderate Expert Panel at SINET Showcase on Bias in Artificial Intelligence Securityon October 28, 2019 at 5:20 pm
2019 on the impact of bias in AI-driven security systems. Dr. Srinivas Mukkamala, co-founder and CEO of RiskSense, is a recognized expert on artificial intelligence (AI) and neural networks. He was ...
- Why privacy and artificial intelligence still need some harmonizationon October 28, 2019 at 12:49 pm
Some of the leading issues include potentially making de-identification impossible, algorithmic bias and patient safety or treatment errors. The arguable inability to de-identify data is one of the ...
- Ethics & bias: Artificial intelligence in the spotlighton October 15, 2019 at 2:30 am
It is no secret that Artificial Intelligence is having a big moment ... Managing the accuracy and morality of AI systems is often overlooked, but it is essential that companies combat bias within data ...
- ABTV Crossfire debate: Does bias in artificial intelligence exist?on September 26, 2019 at 8:24 am
The second and the most scientific reason for human bias bleeding into artificial intelligence and other systems is data sets. AI systems are only as good as the data we put into them. Bad data can ...
- Art Project Shows Racial Biases in Artificial Intelligence Systemon September 24, 2019 at 12:04 pm
Some of the labels proposed by ImageNet Roulette—an artificial intelligence classification ... “We want to show how layers of bias and racism and misogyny move from one system to the next,” Paglen ...
- Artificial Intelligence has a gender bias problem – just ask Sirion September 22, 2019 at 2:16 am
Virtual Personal Assistants – such as Bixby, Alexa (Amazon), Siri (Apple) and Cortana (Microsoft) – are at the cutting edge of marketable artificial intelligence (AI). AI refers to using technological ...
- South Africa: Artificial Intelligence Has a Gender Bias Problem - Just Ask Sirion September 21, 2019 at 5:00 pm
In South Africa, where I live and conduct my research into gender biases in artificial intelligence ... I picked up in engineering." Bias and discrimination in AI It is being increasingly acknowledged ...
- How Bias Distorts AI (Artificial Intelligence)on August 5, 2019 at 7:35 pm
When it comes to AI (Artificial Intelligence), there’s usually a major focus ... Can you share some real-life examples of bias in AI systems and explain how it gets there? It’s a common misconception ...
- How Bias Distorts AI (Artificial Intelligence)on August 4, 2019 at 7:38 am
When it comes to AI (Artificial Intelligence), there's usually a major focus on ... So here's a look at what she has to say about bias: Can you share some real-life examples of bias in AI systems and ...
via Bing News