Dual Use of Artificial Intelligence-powered Drug Discovery
By Fabio Urbina, Filippa Lentzos, Cédric Invernizzi, Sean Ekins. Nature Machine Intelligence, 2022
355 citations recorded by the source corpus.
An international security conference explored how artificial intelligence (AI) technologies for drug discovery could be misused for de novo design of biochemical weapons. A thought experiment evolved into a computational proof.
Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology
View Paper: https://www.semanticscholar.org/paper/d7eef9b5bb65feda6647440e7727bbcdf0edaebc
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Theory and Mathematics
- Seyone Chithrananda
- Gabriel Grand
- Philippe Schwaller
- Mantas Mazeika
- Richard Ngo
- Bharath Ramsundar
- Jimmy Ba
- Yan Shoshitaishvili
In Materials Chemistry
- Seyone Chithrananda
- Gabriel Grand
- Philippe Schwaller
- Bharath Ramsundar
- Jimmy Ba
- Kevin M. Esvelt
- Yoshua Bengio
- Yang Liu
In Molecular Biology
- Seyone Chithrananda
- Gabriel Grand
- Philippe Schwaller
- Mantas Mazeika
- Elana Simon
- Bharath Ramsundar
- Jimmy Ba
- Dan Hendrycks
Citations in the corpus
- The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning (2024, result): “Similar to other technologies, such as gene editing and nuclear energy, AI is dual-use —it can be leveraged for benefit and harm (Urbina et al., 2022).”
- ChemBERTa-2: Towards Chemical Foundation Models (2022, background): “While recent work has highlighted the dual use risk [22] of chemical models, for now we believe that the challenge of synthesizing a novel molecule limits the potential harms from releasing updated models.”
- Augmenting large language models with chemistry tools (2023, background): “Nevertheless, it is crucial to ensure responsible development and use of LLM agents 66–68 .”
- Empowering Biomedical Discovery with AI Agents (2024, neutral): “Responsible development of agents requires developing preventive measures [86, 87].”
- GPT (Generative Pre-Trained Transformer)— A Comprehensive Review on Enabling Technologies, Potential Applications, Emerging Challenges, and Future Directions (2023, background): “GPTs are capable of learning new patterns and relationships [89] in the dataset they were trained on.”
- The alignment problem from a deep learning perspective (2022, background): “An early example of AI weapons development capabilities comes from an AI used for drug development, which was repurposed to design toxins [Urbina et al., 2022].”
- An Overview of Catastrophic AI Risks (2023, background): “In 2022, researchers took an AI system designed to create new drugs by generating non-toxic, therapeutic molecules and tweaked it to reward, rather than penalize, toxicity [11].”