Building Privacy-Preserving and Secure Geospatial Artificial Intelligence Foundation Models
arXiv:2309.17319 · doi:10.1145/3589132.3625611
Abstract
In recent years we have seen substantial advances in foundation models for artificial intelligence, including language, vision, and multimodal models. Recent studies have highlighted the potential of using foundation models in geospatial artificial intelligence, known as GeoAI Foundation Models, for geographic question answering, remote sensing image understanding, map generation, and location-based services, among others. However, the development and application of GeoAI foundation models can pose serious privacy and security risks, which have not been fully discussed or addressed to date. This paper introduces the potential privacy and security risks throughout the lifecycle of GeoAI foundation models and proposes a comprehensive blueprint for research directions and preventative and control strategies. Through this vision paper, we hope to draw the attention of researchers and policymakers in geospatial domains to these privacy and security risks inherent in GeoAI foundation models and advocate for the development of privacy-preserving and secure GeoAI foundation models.
arXiv comment1 figure
References in corpus (29)
- LLaMA: Open and Efficient Foundation Language Models
- Language Models are Few-Shot Learners
- On the Opportunities and Risks of Foundation Models
- Evaluating Large Language Models Trained on Code
- BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models
- Segment Anything
- Toolformer: Language Models Can Teach Themselves to Use Tools
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Constitutional AI: Harmlessness from AI Feedback
- Extracting Training Data from Large Language Models
- HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face
- Ignore Previous Prompt: Attack Techniques For Language Models
- On the Opportunities and Challenges of Foundation Models for Geospatial Artificial Intelligence
- LSTM-TrajGAN: A Deep Learning Approach to Trajectory Privacy Protection
- Adaptive Reward-Poisoning Attacks against Reinforcement Learning
- Reconstructing Training Data from Trained Neural Networks
- Generative AI Meets Future Cities: Towards an Era of Autonomous Urban Intelligence
- BadGPT: Exploring Security Vulnerabilities of ChatGPT via Backdoor Attacks to InstructGPT
- Federated Foundation Models: Privacy-Preserving and Collaborative Learning for Large Models
- AD-AutoGPT: An Autonomous GPT for Alzheimer's Disease Infodemiology
- CSP: Self-Supervised Contrastive Spatial Pre-Training for Geospatial-Visual Representations
- ImageBind: One Embedding Space To Bind Them All
- Instructions as Backdoors: Backdoor Vulnerabilities of Instruction Tuning for Large Language Models
- Towards Building the Federated GPT: Federated Instruction Tuning
- Evaluation and Enhancement of Semantic Grounding in Large Vision-Language Models
- Philosophical Foundations of GeoAI: Exploring Sustainability, Diversity, and Bias in GeoAI and Spatial Data Science
- Bag of Tricks for Training Data Extraction from Language Models
- Autonomous GIS: the next-generation AI-powered GIS
- CATS: Conditional Adversarial Trajectory Synthesis for Privacy-Preserving Trajectory Data Publication Using Deep Learning Approaches