'Major Technical Challenges' Hinder Path to Responsible AI, NIST Official Says

Published on March 9, 2023

New artificial intelligence tools that interact with human users - often powered by large language models - visibly lack federal guardrails. The National Institute of Standards and Technology’s (NIST) Elham Tabassi takes stock of the ‘technical challenges’ in the way of achieving broadly responsible AI.

NIST, which recently released its new AI Risk Management Framework, is paying particular attention to the technical frameworks being used to develop AI tools. Because these tools are intended to become widely available for public use, their manipulation of data must be critically assessed by developers from the outset. This, Tabassi notes, is in the interest of a ‘common ground’ on trustworthy AI.

Read the full article here.

Other news

Defence signs on to university-led artificial intelligence research

March 9, 2023
The Australian Department of Defence commits a total of $1.746 million to fund two artificial intelligence research projects. The projects will be carried out by the University of South Australia and Deakin University.The goal of the projects is to learn how ‘to process noisy and dynamic data into information and help inform military decision making.’ The funding is intended to rapidly establish t

Center of Excellence Opens in Tawazun Industrial Park, Serving UAE’s Autonomous Systems Community

March 9, 2023
The United Arab Emirates’ Tawazun Council signs an agreement with Boeing to establish a Center of Excellence for uncrewed systems at Tawazun Industrial Park. The Center will work to support the UAE Armed Forces as it expands its modernization efforts.The facility will conduct aircraft repairs and overhauls, in addition to serving as a training hub for new operators. The workforce will also ‘develo