Case History on Embracing Opportunities of Disruptive Technologies and Creating Trust in AI Systems
Event Details
Abstract: Generative Artificial Intelligence/Machine Learning including Generative Pre-trained Transformers (GPT) present both challenges and opportunities that must be addressed by individuals and organizations. Key Take-Aways: Insight on opportunities in generative AI/ML systems
Event Details
Abstract:
Generative Artificial Intelligence/Machine Learning including Generative Pre-trained Transformers (GPT) present both challenges and opportunities that must be addressed by individuals and organizations.
Key Take-Aways:
- Insight on opportunities in generative AI/ML systems to drive rapid knowledge discovery and innovation
- An understanding of the challenges that must be addressed to ensure responsible adoption; workforce disruption, privacy risks, intellectual property leaks, and insufficient transparency and explainability
- A guide for moving forward where individuals and organizations should implement policies, processes, and technologies to facilitate proper human-machine engagement and control during human-machine teaming; start with prioritizing broad education and training to promote AI/ML fluency
Some References for Further Reading:
2016 Defense Science Board Summer Study on Autonomy
National Security Commission on AI
DoD Policy 3000.09 – Autonomy in Weapon Systems
DoD Responsible AI Strategy and Implementation:
- Responsible (judgement/care)
- Equitable (un biased, e.g., early biased systems in resume selection, sentencing, loans)
- Traceable (transparent, explainable)
- Reliable (v&v, testing something that learns/changes, certification and accreditation)
- Governable (detect and avoid unintended consequences)
EU AI Act (presentation here). Regulating use cases: complexity, opacity, unpredictability, autonomy, data which map onto safety, rights, enforcement, uncertainty, mistrust, and fragmentation. Codes of conduct.
NIST AI Risk Management Framework (RMF)
MITRE ATLAS™ (Adversarial Threat Landscape for Artificial-Intelligence Systems), is a knowledge base of adversary tactics, techniques, and case studies for machine learning (ML) systems based
more
Time
(Wednesday) 1:00 pm - 2:00 pm , Eastern Daylight Time
Location
Remote, Also To Be Available On-Demand