Publishing content and copyright debates in the training of artificial intelligence models

Halil Çelik, President of the Press and Publishing Union Association, stated that the unauthorized use of publishing content in the training of artificial intelligence models can be transformed into a new copyright and licensing area.

Unauthorized content use and labor

Halil Çelik, President of the Press and Publishing Union Association, stated that the unauthorized and uncompensated use of publishing content in the training processes of artificial intelligence systems poses a major problem.

Emphasizing that authors, translators, editors, publishers, and designers put great effort and economic investment into creating a book, Çelik underlined that this labor must not go unrewarded.

Commercial models and the transparency issue

It was pointed out that content with commercial value created by publishers over the years provides input for the models of artificial intelligence companies.

It is stated that there is still no adequate mechanism to enable publishers to know which book is used in the training of which model and to what extent.

Cultural heritage and rare works

The possibility of old and rare books being digitized to be used in the data sets of artificial intelligence companies increases sectoral concerns.

It is stated that rare works may be physically torn apart and destroyed during the scanning process, which would harm cultural heritage.

Collective licensing and copyright solutions

Artificial intelligence companies are expected to be transparent and make disclosures regarding which content they use for training.

It is strongly emphasized that collective licensing and copyright payment models must be established for commercial artificial intelligence training.