No tags yet.
Diversity in Artificial Intelligence
The impact of a lack of diversity in Artificial Intelligence may very well hinder the practicality of emerging technologies.
Simply put, there aren’t enough women and people of color involved in the development of Artificial Intelligence and technologies which use Artificial Intelligence. This can have disastrous effects on the serviceability of AI.
As Princeton University professor, Dr. Olga Russakovsky, explained in a presentation about inclusion in computer vision, a lack of diversity and bias leads to an endless cycle of discrimination in technology. Essentially, AI is a technology which is meant to benefit and reflect all of its users. However, it is currently being built by a homogenous group of people, primarily white males, who unintentionally bring in their own personal biases and prejudices which are later reflected in the large datasets used in Artificial Intelligence technology. These datasets are used to train the AI models which then make biased decisions based solely off of the data sets they have been subjected to. These decisions further perpetuate the lack of diversity within AI and thus the cycle continues.
Computer Vision Meets Fairness | Princeton University
In order for all of its users to be able to reap the benefits of Artificial Intelligence, its authors need to be a diverse set of people. Encouraging diversity in this field should start at a young age. As Dr. Pulimood, a computer science professor at TCNJ, explained, girls are often dissuaded from pursuing STEM-related majors from as early as middle school. This leads to a disproportionate number of girls studying computer science after high school. In fact, a 2013 research study found that only 17.9% of students who graduated with a bachelor’s degree in computer science and 19.3% of those who graduated with a bachelor’s degree in engineering were women.
Artificial Intelligence is seeping into every corner of our lives. It’s being used by the little robots which clean our living rooms, the machines that process ER patients’ symptoms to churn out diagnoses, and even major social media companies to help determine which ads should appear in our feeds. No longer can we remain passive about the great gender gap in artificial intelligence and other technologies; time is running out.
We must ask ourselves one pertinent question: what can I, a netizen in 2019, do to support diversity and inclusion in AI?
-Aparna
Original Cover Image: Geeta Lalvani