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
- Artificial Intelligence: The Next Generation Anti-Corruption Technologyon August 24, 2021 at 12:30 pm
Artificial Intelligence is an extremely powerful weapon to use in any kind of sector. Now it helps curb corruption by developing an anti-corruption technology.
- Europe’s artificial intelligence blindspot: Raceon August 23, 2021 at 6:00 pm
Europe's vision of artificial intelligence regulation is color-blind ... referring to a slew of legal complaints in the U.S. brought forward as a result of racial bias in AI systems. "In Europe we ...
- I rest my case, Honourable Algorithm: Artificial intelligence could help eliminate judicial biason August 23, 2021 at 12:26 pm
Studies in the US have demonstrated that artificial intelligence has been proven to predict outcomes of court cases better than lawyers. There is also an indication to suggest that AI could make ...
- MASS SEEKS TO CREATE ARTIFICIAL INTELLIGENCE TASK FORCE, WON’T PUBLICLY NAME APPOINTEESon August 18, 2021 at 6:47 pm
This body does not seem to be centering issues related to accountability, justice, racial discrimination or other forms of bias.” ...
- HIMSS21: How Artificial Intelligence Can Improve Patient Outcomeson August 17, 2021 at 6:23 am
AI can be used to identify population health trends and make predictions to better patient care. How else can this technology transform healthcare?
- Time for responsible Artificial Intelligenceon August 14, 2021 at 5:01 pm
Amazon, in 2018, closed down an AI-based recruiting tool after the e-commerce major figured the tool was biased against women ...
- NCC Conversations: How can we reduce bias in artificial intelligence?on August 3, 2021 at 9:56 am
However, with this growing influence comes concerns about making artificial intelligence systems more inclusive and accessible. To build a safer and more secure future for all, minimising bias in ...
- Hybrid hiring: How artificial intelligence AND human intuition helps counter hiring biason July 26, 2021 at 7:01 am
David Bernard, CEO of predictive recruitment platform AssessFirst evaluates how a merged method using AI and algorithmic assessments can help predict the perfect candidate and remove ‘intuition bias .
- Ethical Artificial Intelligence: Potential Standards for Medical Device Manufacturerson July 19, 2021 at 5:00 pm
While artificial ... Systems (ECPAIS) in 2018. The purpose of the program is to develop specifications and a certification framework for developers of AI systems who work to mitigate issues of ...
via Bing News