Who is the author: the question in the age of artificial intelligence
When a system based on artificial intelligence creates material or fully determines it, the first legal question concerns authorship: does copyright arise in such material at all, and to whom does it belong? In Georgia this question is governed by the Law of Georgia on Copyright and Neighboring Rights. The law contains no separate norm on artificial intelligence, so the answer must be constructed from its foundational concepts: what is an object of copyright, when is a work deemed created, and who is deemed an author. Below we explain these concepts and show where the case of material produced by artificial intelligence creates a distinctive difficulty.
What is an object of copyright
Under the law, copyright extends to works of science, literature and art that are the result of intellectual-creative activity — irrespective of the purpose of the work, its merit, genre, volume, form or means of expression. Two features are decisive: first, that the work is the product of creative activity, and second, that it exists in an objective form. The law draws the boundary expressly: copyright does not extend to ideas, methods, processes, systems, means, concepts, principles, discoveries or facts — even if they are expressed, described, explained, illustrated or embodied in a work. This norm matters greatly in the practice of artificial intelligence: the text of an instruction, an idea or a request that a user gives to a system is not an object of copyright unless it has been shaped into a separate work. The output side must be examined separately — the question is always whether the result carries the trace of creative activity fixed in an objective form.
Categories of protected works and derivative works
The law enumerates the categories of works: literary works — a book, brochure, article, computer program and others; dramatic and musical-dramatic works, choreographic works and other stage works; musical works with or without text; audiovisual works; works of sculpture, painting, graphics, lithography and other fine art; decorative-applied and monumental art; theatre-design works; architectural, urban-planning and landscape-garden works; photographic works and works created by means analogous to photography — a still frame of an audiovisual work not being deemed a photographic work; maps, plans, sketches and illustrations relating to geology, geography, cartography and other fields; derivative works, namely translations, screenplays of literary works, adaptations, screen versions, reviews, stagings, compilations, musical arrangements and other transformations; composite works — collections, encyclopedias, anthologies, databases and other works that are the result of creative activity in the selection and arrangement of material; and other works.
For our context the regime of derivative and composite works is especially significant: their protection is exercised regardless of whether the works on which they are based or which they contain are themselves objects of copyright, and the protection of derivative and composite works is exercised equally with the original. This makes the derivative-work frame a natural analytical starting point when human creative contribution and machine generation are combined. The regime of computer programs deserves separate note as well: protection extends to computer programs of every kind, including operating systems, which may be expressed in any language and form, including source text and object code — a program is protected as a literary work.

