Machine-learning generative Artificial Intelligence tools, specifically large-language models, provide varied functionality, like content generation, user-facing chatbots, and code generation. The LLM typically works with a decision engine, such as a neural network. LLMs suffer issues with training data poisoning, copyright of generated content, and this paper's focus; prompt engineering attacks and training data leaks. The authors propose an architecture to co-locate a filtering mechanism with the LLM chatbot to identify and preventing disclosure of leaked LLM training data before communication to the end-user. Implementation of a resource description framework (RDF) based filtering mechanism compares LLM outputs against a bank of training data using three approaches; the first uses a bank of hash-codes generated from training data artifacts, the second uses a bank of training data stored as plaintext, and the third couples natural language processing (NLP) with the plaintext training data bank. Accuracy, overhead and acceleration results are detailed, and observed anomalies in LLM responses to testing including plausible leaks are also discussed.

Public blockchains and distributed ledger technologies (DLTs) provide reliable distributed, decentralized datasharing, with consensus mechanisms and cryptographic security, data integrity, transparency and trust. Though this transparency and open access allows malicious actors to record arbitrary information to the DLT by exploiting free-form editable text fields in request headers or transaction bodies. The authors propose an architecture that runs external agents interfaced with blockchains, capable of filtering unwanted content to mitigate this threat. Implementation of a resource description framework (RDF) based verification mechanism for blockchain input fields, and an alternate approach using natural language processing (NLP) are then described. Finally, the efficiency and effectiveness of the solution is demonstrated on a blockchain deployment.

Blockchains and related applications continue to grow as an area of technology that can provide advantages over traditional approaches in certain use cases. One such area are data markets. Data market use cases include secure asset transactions, data lineage and data governance between multiple parties and involves data security, immutability, transparency. Data markets require strong privacy guarantees for digital asset procurement. The inherent properties of a blockchain can provide significant benefits in the data-market use case. In fact, there are many commercial platforms currently leveraging blockchain technologies for blockchain-based data markets. Leveraging blockchain for data market use cases creates the opportunity for new logical architectures to preserve privacy in the blockchain ecosystem. This publication proposes a multi-faceted solution for stronger privacy guarantees for digital asset procurement on blockchain-based data markets. An architecture will be detailed which provides better guarantees of privacy to those acquiring digital asserts from blockchain-based data markets by leveraging time-delays, batch requests, anonymization, and virtual “one-time” users to form a layer of obfuscation around requests.