Free beta access. 5 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Expanding TEVV is necessary to ensure confidence in AI systems. Continuous testing is better than current efforts.

National Security Commission on Artificial Intelligence 2022 [March 19, “NCSAI Final Report 2021 – Chapter 7: Establishing Justified Confidence in AI Systems”, https://assets.foleon.com/eu-west-2/uploads-7e3kk3/48187/nscai_ch7_digital_02-26-21.24159789cf2e.pdf]

3. Testing and Evaluation, Verification and Validation (TEVV). Having justified confidence in AI systems requires assurances that they will perform as intended, including when interacting with humans and other systems. The TEVV of traditional legacy systems is not sufficient at providing these assurances. As a result, agencies lack common metrics to assess trustworthiness that AI systems will perform as intended. To minimize performance problems and unanticipated outcomes, an entirely new type of TEVV will be needed. This is a priority task, and a challenging one. The federal government will need to increase R&D investments to improve our understanding of how to conduct AI and software-related TEVV. Toward this end: DoD should tailor and develop TEVV policies and capabilities to meet the changes needed for AI as AI-enabled systems grow in number, scope, and complexity in the Department. This should include establishing a TEVV framework and culture that integrates continuous testing; making TEVV tools and capabilities more readily available across DoD; updating or creating live, virtual, and constructive test ranges for AI-enabled systems; and restructuring the processes that underlie requirements for system design, development, and testing.6 National Institute of Standards and Technology (NIST) should provide and regularly refresh a set of standards, performance metrics, and tools for qualified confidence in AI models, data, and training environments, and predicted outcomes. NIST should lead the AI community in establishing these resources, closely engaging with experts and users from industry, academia, and government to ensure their efficacy.

The DOD must improve TEVV by extending testing for the life cycle of the system – this is essential for AI which evolves over time.

Flournoy, Haines et al 2020 - former Under Secretary of Defense for Policy and Director of National Intelligence [Michèle and Avril, Gabrielle Cheftiz Special Assistant to the Under Secretary of Defense for Policy October, “Building Trust through Testing Adapting DOD’s Test & Evaluation, Validation & Verification (TEVV) Enterprise for Machine Learning Systems, including Deep Learning Systems” https://cset. georgetown.edu/wp-content/uploads/Building-Trust-Through-Testing.pdf Acc 6/23/22 JZ]

5. Bridge the gap between development and testing. ML/DL systems will require testing and verification across their entire life cycle, which will require stronger links between program managers and testers as well as methods to capture lessons learned throughout deployment. There are already good models for how this could be done. For example, developers for Project Maven need to submit to T&E in every sprint cycle, or they cannot move forward to the next stage of development. We recommend replicating and scaling this approach in other programs. DoD could also look to the Joint Improvised-Threat Defeat Organization, established in 2006, to rapidly field new technologies to counter IEDs. The organization’s approach to fielding prototypes and testing them in the field provides a useful model for the rapid deployment of ML/DL. Another way to bridge this gap is to leverage the testing framework and requirements language to inform the acquisition process. The JAIC is already using its acquisition process to impact development of AI technologies by using the AI ethical principles as “applicable standards.” One could envision a similar process with AI testing and governance standards. To do so, DoD will need to maintain a robust dialogue with industry on what is feasible to help inform the process.

Sign in to recut, update, or save this card.