digiM Statement on the Use of Artificial Intelligence Technologies

As artificial intelligence (AI) technologies advance and become more prominent in research and development, assuring data integrity, accountability, security and confidentiality is paramount. This disclosure statement describes the principles and practices of digiM’s use of machine learning (ML), deep learning (DL), AI, and generative adversarial network (GAN) technologies, under the guidance of digiM’s core value in trust, honesty, and transparency. 

digiM’s in-house, proprietary, commercially available software incorporates AI-based image processing tools including image denoising, image segmentation, super-resolution, and structure synthesis. These AI tools serve as assisting analysis methods and do not replace the decisions of digiM scientists. All digiM’s decisions and deliverables incorporate human in the loop, and are cross validated whenever necessary to ensure sufficient training data, quality training data, well posed training strategy, and rationalizable prediction results.

digiM computational physics prediction tools, including patented release prediction, patented single phase and multi-phase flow permeability, two phase capillary pressure, molecular diffusion, and dissolution, do not use AI or large language models in their predictions. 

digiM implements all AI tools in house, running on digiM owned proprietary hardware physically hosted in digiM offices. digiM does not use any commercial AI software tools, nor large language models for any deliverables to clients. digiM does not send data to any external, third party servers.

digiM does not risk data confidentiality under any circumstances. The data used to train AI algorithms are strictly limited to the data generated for a specific client. When public data or authorized data are considered beneficial for the client’s project, they will be fully disclosed and approved by the client before usage.

digiM warrants that any use of the AI will not violate any third party rights. digiM further warrants the best effort to comply with applicable laws and then-current generally accepted artificial intelligence industry standards.

‍