AI/Robotics Update A new advisory committee was established by the Biden administration, the National Artificial Intelligence Research Task Force (NAIRRTF). The purpose is to give recommendations to the NAIRR (National AI Research Resource) and to submit reports to Congress regarding overview, security and technical abilities of AI.
There is a movement regarding AI being “trustworthy.” An example of this concern/movement is that The National Institute of Standards (NIST) has recently launched its Trustworthy Responsible AI Resource Center (TRARC). This is to offer a resource to organizations with trustworthy development and use of AI technology. The framework initiated by NIST is referred to as AI RMF 1.0 and is available for review from NIST.
Another aspect of trustworthy AI is being cautioned by insights offered through universities. The basic concern is that current parameters for AI development do not address true trustworthy AI use and development as it focuses too much on a narrow window of use and operation, but rather should consider the entire potential for hazards from a complete overview of all possible interactions and interfaces.
Other concerns for AI and Robotics being safe and trustworthy have been raised with the development and design of mobile and collaborative robotics, especially those utilizing AI programming. The traditional physical safety zone and restricted zones around robotics no longer exist when the robotics are mobile and designed to be autonomous and interactive with humans, other machines, and technology interface scenarios. Thus, how to insure all interactions and operations have safe and fail-safe conditions for all conceivable interactions and operations.
Another concern in the AI algorithm development stage that has been heard about in the news lately is AI being utilized for writing papers and research work. The challenge currently is that AI programming has a “guessing” factor built into the “learning” mode of AI. Thus when writing a paper or being asked for research information the AI could offer incorrect data as the AI can be guessing the answer while it is learning. And the information the AI passes along may not be “reliable” information. Thus, it is being proposed (demanded by some) that proof-checking be required or an updated version of how AI “learns” or researches to assure accuracy in its responses.
The WMMA Technical Standards Committee has been monitoring/discussing information about AI and robotics as it continues to develop due to the potential impacts and implications for woodworking machinery and standards development.
The Robotics Industries Association has these standards for robotics currently available from ANSI; ISO 10218 parts 1 & 2, RIA R15.06 & RIA R15.08.
Email to a Friend