> The model will analyze phone and video calls, text messages, and emails and then flag sections for human agents to review. These agents then send them to investigators for follow-up.
Given that US citizens already pay for adequate mass surveillance programs like XKeyscore and Carnivore, for example, are inmate communication monitoring AI systems unnecessarily redundant?
> The model will analyze phone and video calls, text messages, and emails and then flag sections for human agents to review. These agents then send them to investigators for follow-up.
- plot all mentions of a word, stemmed word, or list of words binned by time interval
- find first occurrence (of a mention, of a repetitive pattern of abuse)
- find common substrings
- find multiple names in sequence; track the count of name references per time interval
- with an LLM, build queries to also search for synonyms and similar phrases
Dystopian, to say the least.
Given that US citizens already pay for adequate mass surveillance programs like XKeyscore and Carnivore, for example, are inmate communication monitoring AI systems unnecessarily redundant?