Mission: MIRI’s mission is to ensure that the creation of smarter-than-human intelligence has a positive impact. We aim to make advanced intelligent machines behave as we intend even in the absence of immediate human supervision.
Programs: Research programin 2014 the machine intelligence research institutes (miri) primary research program goals were to (1) increase its technical research output and (2) invest heavily in recruiting additional technical researchers. To this end miri released 14 technical papers/reports including a brand new technical agenda, gave a few technical talks for academic audiences, ran a research workshop, hired two new full-time research fellows, launched the mirix program with over 10 groups (this exposed dozens of potential researchers to miri research), published a guide to miris research to help potential new researchers understand miris research agenda, and hosted 8 visiting researchers. In 2014 miri continued to deemphasize strategic research, however it still produced 15 strategic analyses posted on miris blog and elsewhere, 54 expert interviews on a wide range of topics, and 5 papers/chapters on ai strategy topics. Outreach programmiri continued to reduce its outreach efforts in 2014 in order to direct more resources to its research program. Nevertheless, in 2014 miri published a new book called smarter than us (a book intended for a general audience which lays out the motivation for miris research), gave 4 talks at the 2014 effective altruism summit and retreat (a conference of over 100 people trying to figure out how to do the most good), hosted a book launch party for nick bostroms superintelligence (a book detailing strategic considerations relevant to miri research which became a new york time best seller), and continued to give interviews to various media outlets.
MIRI is one of the few beacons of research on AI safety, called the control problem. They collaborate with the future of humanity institute and the FLI. This is perhaps the most important research of our times, given the possibility of advanced AI fairly soon(estimates place about 50% probability for the year 2050). I would highly recommend that you support MIRI in their efforts.