Overall then, the AI community has achieved some successes as workers organising and bargaining with their employers. This may be attributed to the organisational products and services (1), organisational production technology (3), and the general economic conditions (4) all being favourable – though the structure of bargaining has not been (2) – and the AI community having been fairly committed to collective action (5). Factors likely to continue include the products and services being consumer-facing and difficult to stockpile; reliance on high-skilled labour; and the unequal bargaining structure. However it is unclear what the balance of talent supply and demand will be, and to what extent the AI community will continue to be committed to collective action. These are key questions for further research.
However I am not that optimistic about unions specifically, because in general they seem mainly focused on benefiting their workers specifically, rather than taking into account impacts on broader society. In the same way that fossil fuel unions, or police unions, or longshoremen unions have interests that significantly diverge from society as a whole, I would expect AI employee unions to still want their employers to aggressively commercialize.
You might find Haydn’s work on the subject interesting:
However I am not that optimistic about unions specifically, because in general they seem mainly focused on benefiting their workers specifically, rather than taking into account impacts on broader society. In the same way that fossil fuel unions, or police unions, or longshoremen unions have interests that significantly diverge from society as a whole, I would expect AI employee unions to still want their employers to aggressively commercialize.