Sean Flynn and Dick Kawooya
Printable version with footnotes (PDF)

We are grateful for the opportunity to participate in Nigeria’s consultation on its copyright bill. Below, we present some of our research findings from PIJIP’s Project on the Right to Research in International Copyright relating to the importance of flexibility in copyright law to permit text and data mining (“TDM”). TDM is a critical element of numerous machine learning and “artificial intelligence” intelligence applications.

Our research supports the adoption of the proposed open fair dealing exception for “research.” Our research also supports consideration of an additional specific exception for uses of works in TDM to supplement the proposed general fair dealing exception.

A. General Exceptions for Research

Empirical research shows that more publication of citable research takes place in countries with more “open” research exceptions — that is, research exceptions that are open to all uses (e.g. reproduction and communication), to all works, and to all users.[1]

The Executive Bill proposes the adoption of an open general exception for research and other purposes in Section 20. Our research supports the adoption of this exception. A general exception is one which covers multiple purposes of use in a single exception.[2] Our research finds that most Commonwealth countries provide a general exception for “fair dealing” with a work for multiple purposes including for “research.” The category of “research” can be interpreted broadly to include uses for consumer research.[3]

Section 20 is an example of what we call an open general exception. By virtue of the addition of the words “such as” before the list of permitted purpose, the Nigerian proposal would operate like the U.S. fair use right and permit application to other purposes. Several countries from both the civil and common law tradition provide open general exceptions.[4] The U.S. fair use right is
one example.[5]

The benefit of an open general exception is that it can accommodate unforeseen uses that are nonetheless fair to the right holder. Countries such as the United States were the first to adopt text and data mining methodologies even before the practice was clearly authorized, thus
gaining significant advantages in the fields of research and technology.

Interestingly, empirical research has shown that transitioning from a fair dealing right with a closed list of purposes to a fair use right with an open list of purposes can benefit research, even where the prior fair dealing right explicitly protected research uses.[6] This may indicate that open exceptions give researchers a positive signal that can be beneficial to their work, even where the fair dealing exception already covers research purposes.

B. Specific Exception for Text and Data Mining

Empirical research evidence shows that text and data mining research is promoted through exceptions that more specifically authorize text and data mining research.[7] Other countries with general research exceptions have added specific exceptions for text and data mining as well.[8] We encourage Nigeria to consider such a specific TDM exception in its law, and present Singapore’s recent copyright amendment as an excellent model.

Many scholars argue that text and data mining should not be considered within copyright’s exclusive protection because copyright was never intended to require authorization for reading and analysis.[9] Nonetheless, copyright questions can be raised with respect to the technical
reproductions required to create a “corpus” of works to be mined for many projects.[10]

Our research indicates that the most useful text and data mining exception is:

  • Open to all TDM “uses,” including specifically to communications or distributions needed to promote research collaboration and validation;
  • Open to all users, both individuals and institutions, commercial and non-commercial;
  • Open to the use of all works, including, for example, audio visual works.

Legislators have defined the purpose of the use protected in TDM exceptions through terms such as “text and data mining”,[11] “computational” use,[12] or “data analysis”.[13]

Singapore has just adopted a very useful and highly specific exception for “computational data analysis” that provides a clear model Nigeria could follow. Singapore’s exception extends to reproductions and communications to the public that are necessary for the purposes of: (i)
verifying the results of the computational data analysis or (ii) collaborative research and study.

The exception encompasses both commercial and non-commercial uses. Article 60 specifically provides that computational data analysis under the exception does not constitute a protected publication. And as the EU Copyright in the Digital Single Market Directive, Singapore provides that computational data analysis “may not be excluded or restricted” by contract.[14]

We would be happy to discuss our research in more detail.you may contact us at sflynn@wcl.american.edu. Again, we are grateful for the opportunity to participate in this consultation.

Respectfully,

Sean Flynn
Director, Program on Information Justice and Intellectual Property
American University Washington College of Law
Principal Investigator, Project on the Right to Research in International Copyright Law

Dick Kawooya
Assistant Professor at the University of South Carolina School of Library and Information Science
Africa Regional Coordinator, Project on the Right to Research in International Copyright Law