WGIC’s comments to the US Department of Commerce on “Addition of Software Specially Designed to Automate the Analysis of Geospatial Imagery to the Export Control”
World Geospatial Industry Council (WGIC) Response to:
Department of Commerce, Bureau of Industry and Security, 15 CFR Part 774
[Docket no. 191217-0116] RIN 0694-AH89
“Addition of Software Specially Designed to Automate the Analysis of Geospatial Imagery to the Export Control”
Classification Number 0Y521 Series
The world we live in is going through turbulent times. Factors like geopolitical volatility, economic instability, and disruptive outbreaks are contributing to global uncertainty, pushing governments all over the world to pay additional attention to national security and public safety. The United States of America, which is inarguably a global leader in cutting-edge technology, is rightfully taking all measures to prevent misuse of technologies that can pose threats to its national security.
The World Geospatial Industry Council (WGIC) (https://wgicouncil.org) is a global association of geospatial companies that endeavours to enhance the role of the geospatial industry and strengthen its contribution to the global economy and society. WGIC performs the following functions:
- Facilitates knowledge exchange;
- Enables creation of larger business opportunities for the industry;
- Represents business interests;
- Undertakes policy advocacy; and
- Maintains dialogue with public authorities, multilateral agencies and other stakeholders.
Being the first, and the only, global organization of geospatial commercial companies to have come together internationally to represent the entire value-chain of the geospatial ecosystem, WGIC understands the U.S. concern, and recognizes that the security and safety of the services, products and tools produced by this sector are essential for national security, and the safety of the customers using these assets.
Geospatial Knowledge Infrastructure in the Fourth Industrial Age
In the global pursuit towards a knowledge-rich society, it is imperative to make actionable knowledge available to users on demand. Knowledge extraction from spatial data is currently the forte of spatial experts, and software to automate geospatial imagery analysis plays an important role in making actionable spatial knowledge readily available to decision-makers, who may or may not be spatial experts. This knowledge infrastructure, and its underpinning tools enhance the efficiency of the overall decision-making process of citizens, organizations and governments.
A geospatial knowledge infrastructure also aims to facilitate this transition through the promotion of technologies and policies which help in making actionable knowledge from spatial data accessible even to non-experts. It also adheres to the agenda set by the United Nations initiative on Global Geospatial Information Management (UN-GGIM) under its Integrated Geospatial Information Framework (IGIF). This Framework will bridge the geospatial divide and promote the use of spatial knowledge to address key global challenges.
Geospatial Knowledge Infrastructure and its Impact on the Economy and Society
The use of geospatial data and knowledge leads to significant benefits for consumers, businesses and governments around the world (see graph below). As per a 2017 report prepared by AlphaBeta for Google (Global Economic Impact of Geospatial Services), geospatial services are estimated to improve the revenues and costs by at least 5% in government services, consumer and retail services, transport, financial services, resources, real estate, utilities, agriculture and food – which all contribute to approximately 75% of the world’s economic output.
In addition to this economic output, geospatial services are also estimated to generate consumer benefits around USD 400 billion. Most of these benefits, however, are concentrated in the developed world, as there exists a major disparity in the access to geospatial services and knowledge between developed and developing countries.
Geospatial Data Key Enabler for Sustainable Development
The United Nations, under the 2030 Agenda for Sustainable Development, through its 17 Sustainable Development Goals (SDGs), pledges to strengthen the capacity of statistical and geospatial information systems in developing countries to ensure access to high quality, timely, reliable and disaggregated data, in order to capitalize on the vast amounts of data available, and reap substantial benefits from it.
Geospatial Infrastructure for Sustainable Development
The transformation model for the geospatial knowledge infrastructure has to involve collaboration between different stakeholders — government-government, government-private organizations, government-academia and government-funding agencies — in order to integrate and effectively use resources from different stakeholders. The transformation of a geospatial knowledge infrastructure is essential for countries to meet the SDGs. The statistical agencies of the countries are responsible for planning and monitoring the SDGs, and hence it is important that the statistical data is integrated with geospatial data in order to extract valuable insights, with an added spatial layer, for monitoring progress on the SDGs. The progress on SDGs can also be tracked and analysed real-time by using Artificial Intelligence (AI) and Machine Learning (ML) methods, which leads to dynamic decision-making.
Role and Relevance of AI in Geospatial Knowledge Infrastructures
With the advancements in AI and Deep Learning, automation of geospatial imagery analysis is easier and more accurate than ever, and contributes to significant gains in productivity and efficiency among both spatial experts and non-experts. AI and ML techniques can be used to process large quantities of data to extract meaningful insights, and build models based on prediction algorithms. With the rapid evolution of Deep Learning techniques which include image classification, object detection, semantic and instance segmentation, knowledge can be extracted from raw geospatial data without the intervention of spatial experts, and can be made available to a multitude of stakeholders for decision-making.
The Deep learning methods in association with spatial analysis will help significantly in faster processing of data and development of spatio-temporal predictive models, which will help stakeholders from different backgrounds like governance, emergency services, businesses, etc. to make decisions more effectively. With the availability of geospatial Big Data, due to the emergence of low-cost nano-satellites and satellite constellations, and connected devices and sensors across the world, it is imperative that the data be used effectively to draw relevant information.
Restriction on Geospatial Imagery Software Export: Impact on Global Market
At a time when the world economy is uncertain, the U.S.’s restriction on the export of AI software for processing geospatial imagery under the 2018 Export Control Reform Act (ECRA) will create ripples in the global market, and lead to a trade imbalance. American companies that sell the imagery analytics software to international enterprises and governments will either have to backtrack on their existing trade agreements, or reframe them after applying for a license, which will be a time-consuming process. It will also hamper the supply chain resulting in further additional losses.
U.S. companies will also need to reorient their strategy with a focus on markets in the U.S. and Canada. This will create a vacuum in the market which will thwart the process of innovation and adoption of 4IR technologies globally.
The use of AI and ML for integration and analysis of data from multiple sources – space-based, airborne, in-situ (on land, in the ocean and in the atmosphere), from people, social media and a multitude of other sources — has revolutionized the relevance and uptake of geospatial technologies to society. At this stage, a rule like this will prove counterproductive, slowing the pace of societal and business growth.
Impact on SDGs
The SDGs are crucial for making the world a better place, and for ensuring safety, security, liveability and equity around the world. Delivering the SDGs, the Sendai Framework on Disasters and numerous other international agreements, as acknowledged by UN-GGIM, the World Bank and many other global organizations, requires the optimal use of geospatial information, and its associated technologies.
The new rule would mean that several countries and developmental organizations relying on these technologies and software will either have to procure them from other sources, look for possible exemptions, or, in the worst case scenario, halt the ongoing work on SDGs altogether. This is detrimental to the free and open sharing of geospatial knowledge for humanitarian and developmental purposes, and also affects the commitment made by UN-GGIM members, including the U.S, to deploy geospatial technologies for achieving the SDGs.
Impact on U.S. Companies
The geospatial industry is both broad and rapidly expanding globally (see graph below). While slightly more than one third of the market share (34.5% in 2017) of this industry is in North America, growth is expected to slow down in the region and in Europe in the near future. This rule will further contribute to the decline in growth and will have other impacts such as:
- It will adversely impact U.S. competitiveness operating outside America;
- It will create a competitive disadvantage for U.S. companies operating outside the U.S. in terms of their market presence;
- It will pose a major challenge for such companies to fulfill their commitments to deliver upgraded versions of technology to their customers outside the U.S., since the latest software already has built-in AI capabilities; and
- It will affect the Artificial Intelligence sector in the U.S. which thrives on the exchange of research and commercial programs across borders.
Even though the WGIC appreciates the security considerations that appear to be the basis for this ruling, the definitions in this rule are so specific that they appear to single out an individual system. This may be narrowly useful, but its impact to the larger suite of U.S. systems, entities and organizations will be detrimental economically, as indicated above, and will most certainly inhibit the fine tradition of global leadership that U.S. companies have had in this field.
The majority of the features, applications and use cases for automating geospatial imagery analysis are based on Deep Convolutional Neural Networks. These are used primarily for commercial and humanitarian purposes such as precision agriculture and disaster response. Additionally, the enforcement of the new rule will not have any impact on the proliferation of the target technology, since the geospatial imagery analysis, features and capabilities mentioned in it have already been embedded in numerous aspects of commercial products.
The rule will incentivize foreign governments to supersede U.S. technology with similar technology procured from other vendors. While the export restriction may have been imposed to prevent geospatial technologies from being used for nefarious purposes, WGIC feels that more harm will come from this ruling, than benefit. We appreciate your attention to our comments, and remain available for any further questions or comments you have.