How prepared are we for deepfakes? Researchers call for shift in AI to protect women | CBC News

🌈 Abstract

The article discusses the growing issue of deepfakes, particularly the use of AI-generated images to create non-consensual nude images of women. It highlights the lack of women in the decision-making process of AI development, which has led to blind spots in considering the potential misuse of the technology. The article also explores legal and technical solutions to address the problem, as well as the need for greater public education and awareness.

🙋 Q&A

[01] Deepfakes and the Lack of Women in AI

1. What are the key issues surrounding deepfakes discussed in the article?

• The article discusses the proliferation of websites and tools that use AI to create non-consensual nude images of women, including cases involving underage girls
• It highlights the lack of women in the decision-making process of AI development as a contributing factor to the problem, leading to blind spots in considering the potential misuse of the technology

2. What are the legal and technical challenges in addressing the issue of deepfakes?

• In Quebec and Canada, there are legal grey zones around the criminalization of producing and distributing deepfake nude images, with the onus often falling on the victim to prove harm and identify the responsible party
• Technical solutions like banning websites or making images unusable by AI models are limited, as users can find ways to circumvent such measures
• The root of the problem is how people choose to use the available technology, which requires education and awareness-raising efforts

3. What are the proposed solutions to address the issue of deepfakes and the lack of women in AI?

• Increasing gender diversity and representation in the AI decision-making process to better anticipate potential misuses of the technology
• Implementing mandatory ethics training for AI developers to broaden their understanding of how the technology could be used by the public
• Strengthening regulations and public oversight on the use of AI technology, including through public hearings

[02] The Need for Gender Diversity in AI

1. What is the current state of gender diversity in the AI and STEM fields?

• The percentage of women working in the tech field is around 25%, with little progress made in the past two decades
• Gender diversity is crucial for ensuring AI systems are representative of the populations that use them

2. How can gender diversity in AI be improved?

• It's not just about increasing the number of women and non-binary individuals in AI labs, but also creating a culture and atmosphere where they can succeed and be valued members of the team
• Initiatives like the "Affecting Machines" project at Concordia University aim to promote the work of women in the AI and STEM fields

3. What are the benefits of greater gender diversity in AI development?

• Greater gender diversity can help address blind spots in anticipating the potential misuse of AI technology, such as the creation of non-consensual deepfake images targeting women
• Diverse perspectives and experiences can lead to the development of more inclusive and representative AI systems