We here investigate the possibility that the observed collimated outflows i
n luminous infrared galaxies (LIGs) and some Seyfert galaxies can be produc
ed in a starburst (SB) environment. In the former source class, in particul
ar, there seems to be some observational evidence for the presence of nucle
ar SBs in some objects. A nuclear disk can be quickly produced by gas infal
l during star formation in a rotating, stellar cluster. We find that massiv
e nuclear SBs with core disk masses M-d similar to 10(8)-10(9) M. and super
-nova rates nu(SN) similar or equal to 5 x 10(-3)-2 yr(-1) (which are consi
stent with the nu(SN)-values inferred from the observed nonthermal radio po
wer in source candidates) may inject kinetic energies that are high enough
to blow out directed flows from the accreting disk surface, within the SB l
ifetimes. In our models, the acceleration and collimation of the nuclear ou
tflow are provided by magnetic fields anchored into the rotating SB disk. T
he emerging outflow carries a kinetic power that is only a small fraction (
a few percent) of the supernovae energy rate produced in the SB. Based on c
onditions determined from observed outflows and disks, we find that moderat
e disk magnetic fields (greater than or similar to 28 x 10(-4) G) are able
to accelerate the outflows up to the observed terminal velocities (less tha
n or similar to few 100 km s(-1) in the case of the Seyfert galaxies and si
milar to 400-950 km s(-1) in the case of the LIGs). The outflow is produced
within a wind zone in the disk of radius less than or similar to 100 pc in
the LIGs and less than or similar to 10 pc in the Seyferts, with wind mass
loss to disk accretion rate ratios (M)over dot(w)/(M)over dot(d) greater t
han or similar to 0.1 (where (M)over dot(d) similar to 100 M. yr(-1)). The
observation of rotating nuclear disks of gas within similar to few 100 pc s
cales, like that in the prototype LIG, Arp 220 (for which (M)over dot(d) si
milar to 100 M. yr(-1)) and magnetized outflows in Seyfert galaxies and LIG
s (with terminal B similar to 10(-5) G at the kiloparsec scales), provides
some observational support for the magnetocentrifugal disk picture drawn he
re.