Artificial intelligence is also important to men and women

Many experts are concerned, artificial intelligence (AI) can develop deviantly, expressing a narrow vision, biased, biased if the designers and manufacturers of them are still mostly men as The current.

AI bias

According to a 2013 study by the American College of Women, only 26% of computer experts are female, down 9% since 1990. Many experts consider this a serious problem, by other industries. STEM such as biology, chemistry, and technology all have a strong increase in women. For reasons, some people believe that the computer industry is more suitable for men and the culture in this industry is not welcome, even hostile to women.

Programs programmed today will shape the systems of the future and if every designer and programmer has the same worldview, we can get computers with narrow minds. .

A study by Carnegie Mellon University (USA) in 2015 showed that women were less likely to show Google ads for high-paying jobs. The team developed the AdFisher tool to create simulation profiles and run browser tests by surfing the web, collecting data about the impact of small changes in the structure and interests of the displayed content. As a result, male users saw 1,800 high-paying job ads, while female users only saw 300 times.

The number of women in the AI ​​industry is declining.(Photo: Healthunits).

In May 2016, American scientists discovered that a popular crime prediction software tends to be racist. This system often places colored people in high-risk groups and leaves whites in low-risk groups.

In a New York Times article, computer scientist Kate Crawford said the computer industry had "white people" problems. She wrote: "If we are not careful in the way we design and train the machine learning system, we will see the deep-rooted forms of bias built for the future of AI. Like all previous technologies. , AI will reflect the values ​​of its creators, if there is no right policy or solution, we risk building computers with a narrow vision with familiar human prejudices. " .

Fix system error

At the Rework conference on deep learning in London in September 2016 - where women in the minority of 500 delegates, Ms. Nikita Johnson - Rework founder - admitted there was gender inequality in the AI ​​industry.

"If the AI ​​is primarily created by men, it will be able to produce false results" - she told Digital Trends. "The data set should be developed by both men and women to ensure a more comprehensive result of gender issues."

Through the event "Women in smart machines" , Ms. Johnson and her colleagues - mostly women - have highlighted their talents to weak and encourage attendees to find partners, colleagues and female mentor. According to her, it is a necessary step towards the goal of increasing the number of women working in this field.

"One reason for the lack of diversity in the AI ​​community is the lack of models for young women to come to. Therefore, they will lack the motivation to consider AI and computer science as a potential choice for my career, " said Mrs. Johnson.

As with all previous technologies, AI will reflect the value of its creator.

Monitoring algorithms to avoid gender bias, help develop comprehensive systems, and events like Rework help motivate women through connections, but the interest and motivation for them should come sooner. , right from the family. Currently the market has a lot of applications, toys teach children the basics of programming.

Shaona Ghosh - an AI expert at Cambridge University - said on Digital Trends: "First, support, trust and encouragement should come from parents."

According to Ms Shaona Ghosh, there is no simple way to solve the problem of overwhelming men in the field of AI and motivate women to participate more in this industry; But for the best of humanity, we need to do so.

Although Ms. Johnson's efforts, Crawford expert and computer industry researchers are respectable, Ghosh said that the most effective effort will start from home with such encouraging parents. The next system becomes a computer scientist.