Two of my big passions in life are the creation of powerful Artificial General Intelligence, and the promotion of free, open and transparent advanced technology for the good of all sentient beings.  These passions come together in the pursuit of free, open and transparent AI and AGI.  This page gathers together a few recent references that discuss free/open/transparent AI and why it’s so important.

First of all, see this brief H+ Magazine article by AI researcher Bill Hibbard and a few colleagues (including me), which is associated with an online petition.   If you support transparent AI please sign the petition here.   Also, there have been two followup posts in response to feedback to that article: one by Bill, and one by me.

See also Bill Hibbard’s paper on  Open Source AI from the First AGI Conference in 2008.

A lengthy and in-depth discussion of the value of the open-source approach for making AGIs treat humans well as they become extremely intelligent, is in a 2012 article I co-authored with Joel Pitt, Nine Ways to Bias Open-Source AGI Toward Friendliness.

Please check out the open-source AGI project I co-founded in 2008, OpenCog (based on work and code from way before that) — it’s been growing a lot lately.

Also check out OpenAI, an amply funded AI project founded by Elon Musk, Sam Altman, Peter Thiel and other American tech-biz titans.   Note that while OpenAI is starting out with open-source development, they are not (currently) committed to open and transparent development for the long term (see this interview for some discussion on this point).

There are plenty of other great open-source AI projects out there too, and even a handful of open-source AGI projects — I’m not going to list them all here.  But on the AGI side, check out OpenNARS, MicroPsi, AERA, and the (mostly Japan-based) Whole Brain Initiative, to name just a few of many.   GoodAI has also expressed interest in open-sourcing a significant amount of their code.  On the narrow-AI side, major tech firms have recently launched a variety of open source software packages for deep learning, such as Google’s TensorFlow and similar tools from Facebook and Microsoft.

Some theoretical/conceptual articles exploring ideas about the future of AGI, relating closely to free/open/transparent AI concepts, are:

• Open-ended Intelligence, by Weaver and Veitas from the Global Brain Institute
• Superintelligence: Fears, Promises and Potentials, my extensive critique of Nick Bostrom’s fear-stirring about superintelligent AGI, and of his arguments in favor of restricting AGI development to specific closed-off elites
• Infusing Advanced AGIs with Human-Like Value Systems: Two Theses, my discussion of plausible general paths for infusing AGIs with values similar to our own, and what kinds of future AGI value systems they might lead to.   Part of the goal is to articulate a path to the creation of beneficial, ethical AGIs that incorporates near-term open source AGI development.