Magnetic semiconductors hold a very special position in the field of spintronics because they allow the effective manipulations of both charge and spin. This feature is important for devices combining logic functionalities and information storage capabilities. The existing technology to obtain diluted magnetic semiconductors (DMSs) is to dope magnetic elements into traditional semiconductors. So far, the DMSs have attracted much attention, yet it remains a challenge to increasing their Curie temperatures above room temperature, particularly for those III-V-based DMSs. In contrast to the concept of doping magnetic elements into conventional semiconductors to make DMSs, here we propose to introduce non-magnetic elements into originally ferromagnetic metals/alloys to form new species of magnetic semiconductors. To demonstrate this concept, we introduce oxygen into a ferromagnetic amorphous alloy to form semiconducting thin films. All the thin films are deposited on different substrates like Si, SiO2 and quartz glass by magnetron sputtering. The structures of the deposited thin films are characterized by a JEOL transmission electron microscope operated at 200 kV. The optical transparencies of the samples are measured using Jasco V-650 UV-vis spectrophotometer. The photoluminescence spectra of the samples are measured using RM1000 Raman microscope. Electrical properties of the samples are measured using Physical Property Measurement System (PPMS-9, Quantum Design). Magnetic properties, i.e., magnetic moment-temperature relations, are measured using SQUID-VSM (Quantum Design). With oxygen addition increasing, the amorphous alloy gradually becomes transparent. Accompanied by the opening of bandgap, its electric conduction changes from metal-type to semiconductor-type, indicating that the inclusion of oxygen indeed mediates a metal-semiconductor transition. For different oxygen content, the resistivities of these thin films are changed by about four orders of magnitude. Notably, all of them are ferromagnetic. All the samples show anomalous Hall effect. Furthermore, their magnetoresistance changes from a very small positive value of about 0.09% to a negative value of about -6.3% under an external magnetic field of 6 T. Correspondingly, the amorphous structure of the thin film evolves from a single-phase amorphous alloy to a single-phase amorphous metal oxide. Eventually a p-type CoFeTaBO magnetic semiconductor is developed, and has a Curie temperature above 600 K. The carrier density of this material is ～1020 cm-3. The CoFeTaBO magnetic semiconductor has a direct bandgap of about 2.4 eV. The room-temperature photoluminescence spectra further verify that its optical bandgap is ～2.5 eV. The demonstrations of p-n heterojunctions and electric field control of the room-temperature ferromagnetism in this material reflect its p-type semiconducting character and the intrinsic ferromagnetism modulated by its carrier concentration. Our findings may pave a new way to realizing high Curie temperature magnetic semiconductors with unusual multi-functionalities.