Susan Silbey, Leon and Anne Goldberg Professor of Humanities, Professor of Behavioral and Policy Science, MIT Sloan School of Management
From LSE Business Review
As artificial intelligence (AI) and machine learning techniques increasingly leave engineering laboratories to be deployed as decision-making tools in Human Resources (HR) and related contexts, recognition of and concerns about the potential biases of these tools grows. These tools first learn and then uncritically and mechanically reproduce existing inequalities. Recent research shows that this uncritical reproduction is not a new problem. The same has been happening among human decision-makers, particularly those in the engineering profession. In AI and engineering, the consequences are insidious, but both cases also point toward similar solutions.
Bias in AI
One common form of AI works by training computer algorithms on data sets with hundreds of thousands of cases, events, or persons, with millions of discrete bits of information. Using known outcomes or decisions (what is called the training set) and the range of available variables, AI learns how to use these variables to predict outcomes important to an organisation or any particular inquirer. Once trained by this subset of the data, AI can be used to make decisions for cases where the outcome is not yet known but the input variables are available.
As a manufacturing expert, I help factories become more productive by refining the way they operate. Small improvements over time can lead to big changes in the long term.
But it wasn’t until I arrived home late and exhausted from a trip that I realized I needed to use these same principles at home. Opening the closet to hang up my coat, I found the closet crowded with kids’ sporting equipment and old school projects. Similarly, evenings at home seemed shorter as our after-dinner hours became crammed with homework and activities.
Where had all our time gone? And how did the space in our cabinets and closets disappear? Why had the job of doing the dishes slipped from right after dinner to right before bed? And why was I finding myself more frequently being drawn into a game of “Dish Tetris,” struggling to fit all the dishes into the dishwasher when they used to fit in just fine?
The gender imbalance in STEM fields is extreme. According to a 2010 AAUW report, boys and girls take math and science courses in roughly equal numbers in elementary, middle, and high school, however far fewer women than men pursue these fields in college. According to the National Science Foundation, 29% of all male freshmen planned to major in a STEM field in 2006 compared to 15% of all female freshmen.
Further, while 57% of undergraduate degrees are earned by women, only 12% of computer science degrees are earned by women. By college graduation, men outnumber women in nearly every science and engineering field.
This divide grows worse at the graduate level and is even wider in the workplace. GirlsWhoCode.com states that women make up half the U.S. workforce, yet hold only 25% of the jobs in the technical or computing fields. To quote from the site: “In a room full of 25 engineers, only three will be women.”
MIT is known for its excellence in computer engineering. It also has an outstanding, but lesser-known, music and arts program. On Veterans Day weekend, computer engineering and music will connect on the MIT campus, and the result could be important innovations in the way music is produced and enjoyed. Read More »
Some people think that the first step of innovation is asking for $10 million. We have grown accustomed to the idea that landing that kind of money from a granting agency or a venture fund is a prerequisite to execute on an idea for a new technology or market —or for that matter, even come up with a good idea in the first place.