Artificial Intelligence Journal Editorial Website
About the Reviewing Process
First review stage
Submitted articles are sent to qualified reviewers
for peer review. To see what may be expected from the review report, please
consult the 'For Reviewers' section of the present website. It is AIJ policy to
obtain three reviews (usually) or two reviews (in special circumstances) for
submitted papers. The use of only two reviews was more common a few years ago.
Reviewers are asked to return their review within 6-8 weeks (4-6 weeks for
Research Notes). Actual reviewing time from the author's point of view may be
significantly more, because of both the time it takes to recruit reviewers, and
the additional delays by some of the reviewers.
A few years ago the Journal had some problems that resulted in very long
reviewing time for a number of submissions. These problems have now largely been
overcome through the introduction of more streamlined procedures. The average time to a first decision on a
submitted paper has been well under 100 days for several years now.
The first review stage may result in the article being rejected, accepted, or
given a 'conditional accept'. The latter is an invitation to submit a
significantly revised article which will be subjected to a second round of
review.
Second review stage
If the article is given a 'conditional accept' then
the author has the option of revising the article and resubmitting it, and it
will be considered as the same article from the administrative point of view.
The revised article should be accompanied by a memorandum where the author
explains whether and how she/he has acted on the suggestions by the reviewers.
It is of course not obligatory to meet all the reviewers' requests, but then an
explanation of why this has not be done must be supplied.
Even if an article has been rejected, it may sometimes be worthwhile to
submit a new article that is based on similar contents as the first one, but
where improvements have been made based on reviewers' comments as well as in
other ways. Such submissions will be registered as new submissions from our
point of view.
An article can normally not obtain 'conditional accept' more than once, so
the reviewing of a revised article will normally result in either an accept or a
reject. Exceptions to this rule are rare.
Anonymity
Authors will not know the identity of the reviewers, unless
the reviewer herself chooses to divulge their identity. Reviewers know the
identity of the authors. Reviewers normally get to see the other reviews of the
same article after they have returned their own review, but will normally not
know the identity of the other reviewers. Exceptions to the last rule are
sometimes made in case of strongly conflicting reviews, where the reviewers may
be invited to interact in order to find out whether this leads one of them to
adopt the other one's position on the paper.