Excellent post, but I'm surprised, browsing through the comments, that nobody has mentioned what seems to me like the obvious trade-off: cash $$$$.

Either you will have to pay money to publish your article (if the journal is open access, and your article is accepted), or you'll have to refrain from publishing the article elsewhere (i.e., making it available to the public). Otherwise, how would the journals make any money? But due diligence... these are the journals you mentioned, with their qualities w/r/t open access or author fees:

IEEE Intelligent Systems -- not open access, http://www.computer.org/portal/web/csdl/abs/mags/ex/2011/01/mex201101toc.htm

Minds and Machines -- not open access, http://www.springerlink.com/content/p16821r7663k/

Ethics and Information Technology -- not open access, http://www.springerlink.com/content/j77j724u4784/

AI & Society -- not open access, http://www.springerlink.com/content/h732536p1k16/

Journal of Artificial General Intelligence -- open access and seemingly no author fees! may be the way to go... http://journal.agi-network.org/

International Journal of Machine Consciousness -- not open access, http://www.worldscinet.com/ijmc/

Artificial Intelligence -- not open access... do not be fooled by the "open access options available", you will have to pay, http://www.elsevier.com/wps/find/journaldescription.cws_home/505601/description#description

Cognitive Systems Research -- not open access, http://www.elsevier.com/wps/find/journaldescription.cws_home/620288/description#description

Topics in Cognitive Science -- looks like an odd journal, "there will be no such thing as an unsolicited topiCS paper. If you have an idea for a topic then submit a proposal for a topic via the topiCS Editorial Manager website. ... to soliciting papers that we know fit the charter of the journal from researchers who we know have something to say. "

AI Magazine -- not open access, http://www.aaai.org/ojs/index.php/aimagazine/issue/view/192/showToc

IEEE Intelligent Systems -- not open access, http://www.computer.org/portal/web/csdl/abs/mags/ex/2011/01/mex201101toc.htm

IEEE Transactions on Pattern Analysis and Machine Intelligence -- not open access, http://www.computer.org/portal/web/csdl/abs/trans/tp/2011/04/ttp201104toc.htm

Autonomous Agents and Multi-Agent Systems -- not open access, http://www.springer.com/computer/ai/journal/10458

Ethics -- not open access, http://www.jstor.org/action/showPublication?journalCode=ethics

Utilitas -- seems to be open access! http://journals.cambridge.org/action/displayIssue?jid=UTI&tab=currentissue

Philosophy and Public Affairs -- not open access, http://onlinelibrary.wiley.com/journal/10.1111/%28ISSN%291088-4963

By the way, the easiest way to tell if a journal is open access is to try to read some of the recent articles.

I'd suggest writing the papers up using standard terminology as lukeprog suggested--I agree with his assessment of the lay-lucidity of Eliezer's CEV and TDT papers, although his other writing is clearly really good. And then I would also submit to http://philpapers.org/. And then maybe submit to one of the few open access journals above or elsewhere.

Maybe I'm biased though, because it just makes me sad to think that at an institution where people don't even care about tenure, people would still be worrying about where to submit papers. Sometimes I think that every time somebody strategizes over which journal submission will lead to the most prestige, somewhere, somehow, a kitten dies. That was never supposed to be the point.