There are surprisingly few customer-facing tasks that don’t require accuracy and truthfulness. LLM-style AI cannot guarantee either. So the only AI-enshittification options that are really available are at organizations who don’t give a shit about their customers (e.g., Republican-led governments), fraudsters, or the replacement of humans in doing unimportant tasks (which a sensible business would just stop doing altogether).
There are surprisingly few customer-facing tasks that don’t require accuracy and truthfulness. LLM-style AI cannot guarantee either. So the only AI-enshittification options that are really available are at organizations who don’t give a shit about their customers (e.g., Republican-led governments), fraudsters, or the replacement of humans in doing unimportant tasks (which a sensible business would just stop doing altogether).