The All-Seeing AI? - securing sexuality podcast episode 79

One of the top questions listeners ask the Securing Sexuality Home Team is how to minimize the amount of data they feed to "the machine." Unfortunately, we don't have a good answer for you, especially now that AI seems to be developing...gaydar? Tune in this week as Wolf and Stef discuss the subtle, even subliminal data points that Large Language Models are using to figure out your most personal information- whether you intend to share this info or not. ​

Securing Sexuality is the podcast and conference promoting sex positive, science based, and secure interpersonal relationships. We give people tips for safer sex in a digital age. We help sextech innovators and toy designers produce safer products. And we educate mental health and medical professionals on these topics so they can better advise their clients. Securing Sexuality provides sex therapists with continuing education (CEs) for AASECT, SSTAR, and SASH around cyber sexuality and social media, and more.


Links from this week's episode:

• Gaydar: Facebook friendships expose sexual orientation https://doi.org/10.5210/fm.v14i10.2611
• Phuong, D., & Phuong, T. (2013). Gender Prediction Using Browsing History, 271-283. https://doi.org/10.1007/978-3-319-02741-8_24
• Wang, Y., & Kosinski, M. (2018). Deep Neural Networks Are More Accurate Than Humans at Detecting Sexual Orientation From Facial Images. Journal of Personality and Social Psychology, 114, 246–257. https://doi.org/10.1037/pspa0000098
• Leuner, J. (2019). A Replication Study: Machine Learning Models Are Capable of Predicting Sexual Orientation From Facial Images. ArXiv, abs/1902.10739.
• Clemens, B., Lefort-Besnard, J., Ritter, C., Smith, E., Votinov, M., Derntl, B., Habel, U., & Bzdok, D. (2022). Accurate machine learning prediction of sexual orientation based on brain morphology and intrinsic functional connectivity.. Cerebral cortex. https://doi.org/10.1093/cercor/bhac323
• Bucher, B., Kobayashi, M., & Watanabe, K. (2023). Do Fictive Sexual Orientations Induce In-Group Bias in Emotion Recognition?. 2023 27th International Computer Science and Engineering Conference (ICSEC), 220-225. https://doi.org/10.1109/ICSEC59635.2023.10329647
• Lee, H., Yang, Y., Davier, T., Forlizzi, J., & Das, S. (2023). Deepfakes, Phrenology, Surveillance, and More! A Taxonomy of AI Privacy Risks. ArXiv, abs/2310.07879. https://doi.org/10.48550/arXiv.2310.07879

​Artificial Intelligence and Personal Privacy: Analyzing the Risks of AI in Inferring Intimate Information

In today's digital age, the rapid advancements in technology have given rise to various concerns regarding privacy. The intersection of privacy and artificial intelligence (AI) has become a pressing issue, as AI systems increasingly rely on personal information to deliver tailored experiences. This article explores the challenges and potential solutions for protecting personal information in the age of AI, highlighting the importance of striking a balance between the benefits of AI and safeguarding individual privacy. 

1. The Role of Artificial Intelligence in Today's Society: Artificial intelligence has revolutionized various sectors, including healthcare, finance, and entertainment. AI systems can analyze vast amounts of data, identify patterns, and make predictions, enabling companies to deliver personalized experiences and improve efficiency. However, the increasing reliance on personal information raises concerns about the potential misuse or unauthorized access to sensitive data. 

2. The Importance of Privacy in the Digital Age: Privacy is a fundamental human right that must be protected, especially in the digital age where personal information is vulnerable to data breaches and unauthorized access. Individuals should have control over their personal data and be informed about how it is collected, used, and shared. Privacy laws and regulations play a crucial role in ensuring individuals' rights are protected and holding organizations accountable for their data practices. 

3. The Challenges of Protecting Personal Information in the age of AI: 

1. Data Collection and Consent: AI systems require access to vast amounts of personal data to function effectively. However, obtaining informed consent from individuals can be challenging due to the complex nature of AI algorithms and the opacity of data processing. Improving transparency and providing clear consent mechanisms are essential for ensuring individuals have control over their personal information. 
2. Data Security: The storage and transmission of personal information are vulnerable to cyberattacks and data breaches. AI systems often rely on cloud-based infrastructure, which introduces additional security risks. Organizations must implement robust security measures, such as encryption and regular security audits, to protect personal data from unauthorized access. 
3. Algorithmic Bias: AI systems can perpetuate existing biases present in the data they are trained on, leading to discriminatory outcomes. Protecting personal information also means addressing algorithmic bias and ensuring that AI systems do not unfairly discriminate against individuals based on their protected characteristics. Organizations must implement rigorous testing and monitoring procedures to mitigate algorithmic bias. 

4. Solutions for Protecting Personal Information in the age of AI: 

1. Privacy by Design: Adopting a privacy-by-design approach ensures that privacy considerations are integrated into the development of AI systems from the outset. Organizations should implement privacy-enhancing technologies, such as differential privacy or federated learning, to minimize the collection and storage of personal data.
2. Improved Data Governance: Organizations must establish robust data governance frameworks that outline clear policies and procedures for data collection, use, and sharing. Implementing strict access controls, data minimization practices, and regular audits can help protect personal information and ensure compliance with privacy regulations. 
3. Enhanced Transparency and Accountability: Increasing transparency around AI systems' data processing practices is crucial for building trust with individuals. Organizations should provide clear and concise privacy policies, explain the purpose of data collection, and offer mechanisms for individuals to exercise their rights, such as data access and deletion requests. 
4. Ethical AI Practices: Addressing algorithmic bias requires organizations to adopt ethical AI practices. This includes diverse and inclusive data sets, algorithmic audits, and ongoing monitoring to identify and rectify bias. Collaborating with external experts and fostering interdisciplinary discussions can contribute to developing ethical AI systems. 

As artificial intelligence becomes increasingly intertwined with our daily lives, protecting personal information is of paramount importance. Striking the right balance between the benefits of AI and privacy is crucial to ensure individuals' rights are respected. By implementing privacy-by-design principles, robust data governance frameworks, and enhancing transparency and accountability, organizations can navigate the intersection of privacy and AI successfully, creating a digital ecosystem that respects and protects personal information in the digital age.



Key Concepts:

• Privacy and security risks
• Artificial intelligence and personal information
• Analyzing online activities and interactions
• Inference of sexual orientation
• Protecting personal identity online
• Facial analysis and predicting sexual orientation
• Privacy regulations and AI laws

Previous Episode Back To Top Next Episode