Professor Timo Honkela's book Rauhankone (~ Algorithms for Peace) takes on a bold task: how could computers prevent misunderstandings and conflicts? Mr Honkela* is a linguist and AI expert, and his basic idea is (in principle) simple: we have already created quite good machine translation systems, but they only concern the "plain information" level in our language exchanges. We should be able to go deeper.
Honkela proposes that using computers and computer networks we could enable direct consultations with very large collections of people: millions or even hundreds of millions. We could then use computers to analyze the "minutes" of such meetings. Moreover, use digital personal assistants in communication and anger control.
Algorithms feature in Harari's Homo Deus, too, but in a broader context. Harari sees a raise of new kind of thinking that is even an alternative to humanism, and calls it dataism. The proponents of dataism see data and algorithms valuable as such, not only because they are useful for humans. Western capitalism defeated socialism (and fascism) not because it was ethical, but because it enabled faster and better data processing. This view is exaggerated but not without merit. Harare even names Aaron Swartz a "martyr of dataism" because he wanted to "liberate" scientific articles i.e. upload them to be freely accessible.
* I haven't met Mr Honkela but apparently my grandmother was his first primary school teacher. Finland is a small place and the town of Kalajoki even smaller.
Monday, December 18, 2017
Subscribe to:
Posts (Atom)
