Challenges of DSIP initiatives

Realising the potential of DSIP involves overcoming several potential barriers. These typically include some combination of a lack of high-quality and affordable infrastructure; a lack of interoperability of standards; organisational resistance and inter-organisational conflicts; a shortage of the skills needed to implement a digital data-driven public sector; and a lack of trust in digital technologies and activities. To elaborate:

·           Data availability and quality: There may be areas with low or irregular data availability, posing a problem if policies and services are relying on data. The data that is available may be of low quality or in an unstructured format that is complicated to process. This can pose a huge challenge for public organisations, which require well-structured data that can be adapted to certain policy classifications. The provision of high-quality data can require significant upfront and follow-up investments before the data can be shared. While data curation is a prerequisite, it may fall short of data quality needs necessary for a given type of re-use. In many cases, a number of complementary resources may be required, ranging from additional meta-data to data models and algorithms for data storage and processing, and even secured IT infrastructures for (shared) data storage, processing, and access.

·           Interoperability: Related to the previous point, ways need to be found to promote a more seamless flow of data between different data sources. This issue has technical (what kind of digital system can be put in place to make existing and new data interoperable?), semantic (metadata and language issues) and governance (how can all stakeholders be aligned to agree upon an interoperability system?) aspects. A specific aspect has to do with the role and effectiveness of data standards, particularly in a mixed ecosystem with legacy and new systems. Box 1 outlines a few recent prominent initiatives to establish international standards and vocabularies for STI. Still, much remains to be done in this area. A question the project could consider is how international efforts related to data documentation and the development of standards for metadata can be consolidated to improve data interoperability?

·           Organisational resistance: DSIP calls for significant inter-agency co-ordination and shared resources, such as standard digital identifiers and a coherent policy framework for data sharing and re-use in the public sector. Yet, public organisations may resist opening-up and sharing their data for a number of reasons. These include siloed thinking and protectionism of data related to conflicting interests among government organisations and individual departments; legal issues in the sense of data ownership; financial issues from the point of view of investments (one organisation setting up the technical and organisational infrastructure for data gathering and initial processing and another acquiring the data for free); and legitimate concerns that data will be misused or poorly interpreted by others who have an inadequate understanding of its meaning and limitations.

·           Investments in skills: Public officials need to acquire appropriate skills to extract meaning from data (i.e. correctly identify the policy problem to address and relevant factors to design potential solutions) and transform this into information and knowledge to inform decision-making. The fast evolution of the DSIP landscape means such skills need to be continually updated. Little information is available on how STI policy organisations are managing the challenge of acquiring digital skills in the context of DSIP.

·           Trust in digital technologies and activities: Personal data are a considerable part of the data processed throughout the public sector, including in DSIP initiatives, where researcher data features prominently. Concerns that DSIP might violate privacy and personal data protection could lead to a lack of trust in digital technologies and act as another barrier to their adoption and use. These concerns could become stronger with the introduction of newer, more advanced technologies and processes, such as machine learning. New concerns are also emerging around automated decision-making, supported by artificial intelligence; data-driven discrimination; new data divides, based on who owns, collects and analyses the data; further entrenchment of inappropriate indicator use and the distortions this can creates that lead to undesirable behaviours; and encouragement of short term perspectives through continuous monitoring.

·           Long-term sustainability of DSIP infrastructures: Like any infrastructure, there are maintenance and use costs that may be much higher than the initial investment costs and difficult to meet. These can go well beyond the costs of maintaining IT systems and include investment in complementary assets, such as digital skills, that take time to build. Furthermore, users sometimes find digital infrastructures underwhelming and refuse to adopt them. This can lead to promising systems being abandoned early on, even before becoming fully operational. Many public sector-led DSIP initiatives seem to be initiated with relatively short-term funding. Developers gamble on them being widely adopted, which would increase the likelihood of them receiving further funding in the future. This has not always happened and some DSIP initiatives have been discontinued. Systems operated by the private sector tend to rely on fees from chargeable services, advertising revenues and venture capital. 


Content Visibility