Quick Thought: Faux-LLMs and the Austerity Gap

A recurring theme in looking at people's negative experiences with LLMs (self-defined 'addiction', self-loathing, heightened social isolation etc.) is the predominance of vulnerable people. Mental health issues, uncertainties about gender and sexuality, victimisation, past traumas... It's only anecdotal (quick, someone do some big research!) but the pipeline from vulnerability to excessive reliance on faux-social LLMs to damaging results and (attempted) withdrawal certainly seems to be a thing.

On the one hand it's entirely predictable, any outlet or perceived support is going to be interesting to those who most need it. But the apathy of the platforms, the chronic indifference to external, real world support mechanisms (including human-led digital ones), the extractive and exploitative design choices users are subject to... these aren't inevitabilities, they're choices. Prominent gaps in social care have been left and chatbots are what increasingly wait at the bottom of them. Platforms with the most nebulous, poorly implemented ideas of care or responsibility imaginable.

It's not just a matter of material or social funding/austerity either, there are overt confessions from platform owners and managers that they have no interest or faith in human led solutions. They're more than happy for recklessly applied tech to swallow up those in need of help. And however much they might try to frame that as 'utopian' I think 'dehumanising' is a far more apt description. They view care as productivity and productivity as slop generation, at least as far as people who aren't them go anyway.

- Dylan