AI Consultation framing.txt

Consultation framing appears to be considering low context to the lower end of human scale context. It still considers humans to be the gold standard. As such the technologies it encompasses are to a large part backward looking, things which are already in place, and have been running in the wild for some time. Hence harms should by now be obvious. The only exception to this is widespread usage of generative AI. However, the mains harms listed against these technologies relate to misinformation, misplaced anthropomorphism, and deception. 
Notes:
Contextual Resolution: Context does not only refer to the depth and breadth of information that a system has available to it, but also the temporal aspects of the available information. I.e. how frequent and how up to date the information is.
Misinformation is a dangerous category to address as it easily removes dissenting voices that should be listened to. Note diversity of thought is about listening to all voices including ones you disagree with. Failure to do this inevitably results in harm.
Deception and manipulation: misrepresentation of a system/entity as another entity. This is clearly a risk, but it is simply an extension of a pre-existing risk extended to new modalities like speech and video. Hence this is not a new technological risk but should be addressed by existing processes to limit misrepresentation. Manipulation is a risk that requires more investigation to understand the overall threat, but it is not new, social engineering has been with us for some time and, in general, attempts to manipulate the public are a prevalent and everyday occurrence.  A common vector for manipulation is advertising and is generally addressed via standards.  
Misplaced anthropomorphism: building relationships, these issues are possible best dealt with via education (Futurama's don't date robots).