Research on the "Gender" of Artificial Intelligence

2019-12-27

If artificial intelligence is a simulation of human intelligence, then who does it simulate and does it have a gender? Whether you believe that gender is a social structure formed by an individual's environment and culture, a biological determinant from an essentialist perspective, or determined by individual differences theory, one thing is undeniable: gender is important. It often affects how we are perceived and what we can do. Gender plays a role in opportunities for certain professions and whether our car navigation systems recognize or ignore our voice commands. In the field of artificial intelligence research, feminized avatars are most commonly used to play the role of virtual assistants. This perpetuates the view that helping roles are best played by women. Earlier this year, at a digital health conference in Melbourne, a medical expert admitted that 25 years ago, when he was still a rural general practitioner, he misdiagnosed a female patient and almost cost her life because he had never seen that particular disease in a female patient before. From this, we can conclude that his operational dataset was biased based on his experiences. Similarly, the biases in artificial intelligence are a result of the inherent biases in our world. These biases exist in the knowledge systems used to acquire professional knowledge, the datasets used to develop predictive models, and various software and hardware. What's worse is that due to the nature of artificial intelligence, these biases become more concealed, especially in deep learning methods that are difficult for humans to explain and understand. In order for AI technology to meet the needs of both men and women, both genders should be the target of innovative solutions, participating in the design of these systems and being represented in datasets and evaluations. For example, unconscious biases should be avoided when determining which features to include or exclude in training predictive models. However, given the current gender gap, this is not a simple task. Globally, women are underrepresented in engineering and information technology classrooms and workplaces, accounting for around 30% in India and even lower in other countries. This results in products and technologies being predominantly designed for men. In AI research, this ratio is close to 10%. Governments, universities, industries, and society as a whole need to work together to establish an ethical framework that harnesses the benefits of artificial intelligence without neglecting issues such as cognitive decline, threats to autonomy, accountability, privacy, security, discrimination, social impact, and economic impact.

RECENT AI NEWS

RECENT AI TOOLS