osint transparency raises ethical questions transparency raises ethical questions [content preview...

© 2018 IHS. No portion of this report may be reproduced, reused, or otherwise distributed in any form without prior written consent, with the exception of any internal client distribution as may be permitted in the license agreement between client and

IHS. Content reproduced or redistributed with IHS permission must display IHS legal notices and attributions of authorship. The

information contained herein is from sources considered reliable but its accuracy and completeness are not warranted, nor are

the opinions and analyses which are based upon it, and to the extent permitted by law, IHS shall not be liable for any errors or

omissions or any loss, damage or expense incurred by reliance on information or any statement contained herein.

of 7

OSINT transparency raises ethical questions

[Content preview – Subscribe to Jane’s Intelligence Review for full article]

Open-source intelligence has shifted from proving its value to becoming an enabler and an

irritant for nations. Allison Puccioni and Melissa Hanham consider the advantages and

disadvantages of the transparency that OSINT affords

Open-source intelligence (OSINT) has developed rapidly since the mid-2000s: new tools include

crowdsourcing, artificial intelligence, ‘deep-learning’ algorithms that discern otherwise imperceptible

patterns in text and images, and platforms to display the data robustly and intuitively. These tools,

and many of the datasets, often align with the burgeoning technological industry, rather than

government and government-subsidised defence industries.

The OSINT concept of deriving finished intelligence from openly available data took form in the early

1990s, when analysts within the United States intelligence community (USIC) sought to harness

commercial technology and growing online datasets as a means of intelligence gathering alongside

longer-established disciplines, such as geospatial intelligence (GEOINT), human intelligence

(HUMINT), imagery intelligence (IMINT), and signals intelligence (SIGINT).






of 7

Airbus Defence and Space imagery showing Humaymim Air Base in Syria. The fitness tracking app Strava released user data in November 2017 that revealed information about its users’ movements, including this trail (bottom inset) around Russia’s main air base in Latakia, Syria. Fitness tracking and geo-enabled social media applications can reveal sensitive information about numbers, locations, and security routes at military installations. (© CNES 2018, Distribution Airbus DS / © 2017 Strava / © 2018 IHS Markit)

1721342

[Continued in full version…]

OSINT, secrets, and governments

Technological advances, skilled practitioners, and new applications may lead to a scenario in which

OSINT will ultimately surpass the conventional apparatus of state tradecraft. Tools, datasets, and

expertise – unhindered by the restrictions and standards of governments – will expand near the

pace of the expansion of the technological industry itself. This could result in the public having an

increased ability to participate in the tradecraft and access to highly granular information about

global activities, patterns, and trends that was once denied to them.

The corpus of openly available data with which to conduct intelligence-grade analysis is growing

rapidly, and it has been appropriated by extra-conventional actors outside the intelligence

community. This proliferation of data and tools for extracting information could create a state of

‘transparency’, in which security- and defence-related activity can be discerned, tracked, and even

predicted at a high rate by the public – and for public consumption – in parallel to traditional

intelligence collection.

How OSINT complements or competes with traditional intelligence analysis remains a legally difficult

question. Those who work in open-source research often do not have access to classified

intelligence reports unless they are leaked, and those who work in traditional intelligence cannot

breach the charter of their security clearances to comment on OSINT reporting and tradecraft. This

dichotomy precludes a thorough evaluation of finished OSINT, and raises many questions about the

most effective and ethical way to use open-source information.

Governments have historically been reluctant to allow access to information that was obtained

through intelligence operations or tradecraft, as well as to the classified tools and systems used to

collect and process this information. Advanced nations regulate sensitive information and collection

systems in a tiered classification system that confines sensitive information to increasingly secured

facilities and restricted-access channels depending on the level of classification.

In the US, according to Executive Order 13526, Classified National Security Information, of 29

December 2009, dissemination of information classified as ‘Top Secret’ can potentially cause “grave

damage to national security” and is restricted to heavily secured Sensitive Compartmented

Information Facilities (SCIFs) using specialised security features to protect such information.

Dissemination of this information is regulated by Title 18 US Code Chapter 115 (Treason, Sedition,

and Subversive Activities Act). Violation of this act – including the unauthorised dissemination of

classified information by intelligence professionals with access to it – is punishable by fines,

imprisonment, or even capital punishment. US law also protects some Israeli interests: in 1997, the

US Kyl-Bingaman Amendment to the 1997 National Defense Authorization Act prevented any US






of 7

commercial imaging satellites from collecting high-resolution satellite imagery over Israel or the

Israeli-occupied territories.

The United Kingdom employs a similar classification system, and classified information is regulated

under the Official Secrets Act of 1989. Furthermore, the UK occasionally employs the Defence and

Security Media Advisory Notice (DSMA- or D-Notice) as a non-binding but powerful means of

influencing the UK media to decline publication for the ostensible sake of increased security.


OSINT advantages

Conventional intelligence is strictly classified and regulated by its originating agency or entity. In the

West, effective intelligence sharing took shape during the Cold War, but public evidence of

intelligence failures continues to suggest an insular nature of traditional intelligence collection, a

strong reluctance to share or disseminate information contrary to the ‘need-to-know’ protocol, and

even a willingness among government policymakers to appropriate disparate elements of

intelligence to support a pre-existing theory or directive.

A series of intelligence shortcomings occurred in the run-up to the Iraq war in 2003, when the US

executive branch’s insistence on the presence of weapons of mass destruction (WMD) in Iraq

contradicted strongly vetted reporting by the United Nations that suggested otherwise. The failure

to accept the verified UN intelligence before the invasion and the subsequent failure to find any

WMD exemplified such issues with traditional classified intelligence. In any contemporary repeat of

an Iraq-type scenario, the OSINT community would generate vast amounts of analysis derived from

commercial satellite imagery, social media, and ground photos.

Some characteristics of OSINT can mitigate some of the pitfalls of conventional classified

intelligence. Openly produced information is modular and can be shared rapidly and freely within

and beyond a government as a means of intelligence, intelligence vetting, or for training purposes.

OSINT articles published in the media may receive millions of views, and the resulting debate can

either strengthen the argument or publicise a contrary opinion.

Open-source information has already been used to debunk erroneous claims made in 2015 by the

National Council of Resistance of Iran (NCRI), an exile opposition group, that Iran had built secret

underground uranium-enrichment facilities in Tehran. The US Department of State was immediately

asked if the NCRI accusation was true, but it declined to comment. Researchers at the Middlebury

Institute of International Studies at Monterey quickly aggregated, analysed, and disseminated

publicly available information that was published by Dr Jeffrey Lewis in Foreign Policy magazine in

approximately two weeks.


Crowd-sourcing solutions

Furthermore, there are many tedious and repetitive tasks that are not yet ready for machine learning,

but which do not require the skill of a seasoned analyst. Outside the intelligence community, large

numbers of enthusiasts or micro-taskers (internet users who complete simple tasks and are paid on

a per-task basis through companies such as Amazon Mechanical Turk) can be brought together to

perform tasks in a form of crowdsourcing that is infeasible in the high-expertise, strictly cleared realm

of the intelligence community.






of 7

Airbus Defence and Space imagery showing the UNESCO World Heritage Site at the Temple of Baal on 29 August 2015. On 30 August 2015, the Islamic State detonated explosives within the structure, destroying the 2,000-year-old multi-denominational site. Commercial satellite imagery enables projects such as Endangered Archaeology in the Middle East and North Africa and the Smithsonian and Iraqi Institute for the Conservation of Antiques and Heritage to monitor archaeological sites in dangerous or remote territory, including regions in the Levant that were under periodic Islamic State occupation. (© CNES 2015, Distribution Airbus DS / © 2018 IHS Markit)

1721343

For example, in the field of biotechnology, a team of researchers at the University of Washington

created a game in 2008 that called on players to fold proteins as part of a research project focusing

on medical and energy applications. This game, known as Foldit, paired human pattern recognition

with machine learning in order to recognise the way in which proteins could fold. Humans can still

perform this work faster than machine-learning tools tasked with the same work, according to a 2010

report in Nature .

Similarly, image crowdsourcing platforms Tomnod and GeoHive – now both part of the commercial

imagery provider DigitalGlobe – ask a large crowd of humans to identify objects in satellite imagery,

often related to crisis and health mapping. Another platform – geo4nonpro.org – provides

commercial satellite imagery of suspected or known WMD sites, and an international crowd

annotates the imagery. Individuals from all walks of life – including chemists, dockworkers,

architects, and area studies experts – are invited to examine and annotate satellite imagery to

identify objects, trends, and patterns, on the premise that expertise should not be limited only to






of 7

imagery analysts. In 2016, two graduate students using this platform found previously unpublicised

evidence of Iran’s missile testing.


OSINT disadvantages

However, the disadvantages of OSINT are significant, and account for some reluctance on the part

of many governments to fully support the capability. Information gathering that may be permissible

in a relatively open society may nevertheless be illegal or forbidden for an intelligence agency

because it may jeopardise civil liberties or put individuals at risk.

This imagery, first published in Jane's Intelligence Review on 15 April 2008, revealed the extent to which China had expanded its naval base on Hainan Island in the South China Sea. Included in the imagery was the identification of a Type 094 SSBN. (DigitalGlobe)

1303544






of 7

Unlike conventional intelligence collection, some private-sector OSINT collection outside formal

national agencies and international security organisations, such as NATO, is not a formalised

tradecraft, is not regulated by the same principles, does not adhere to national-level protocols, and

does not operate under a standardised code of ethics or behaviour. Some non-state OSINT

collectors may or may not be aware of the level of personal or national security risk that they are

taking, and indeed there is some crossover with investigative journalism in many cases; for example,

two Reuters journalists were arrested in Myanmar under Official Secrets legislation in December

2017 for possessing supposedly classified documents. The journalists maintained that the

documents were publicly available and had been planted on them in a sting operation, but were

denied bail on 31 January 2018.

Within the intelligence community, transfer of information within a country or to a second country is

controlled for reasons of national security or to support international agreements on issues such as

nonproliferation and arms control. Commercial satellite imagery is also regulated in the open-source

community, but the efficacy of these regulations is diminishing as more commercial technology

companies – many in nations without the same level of operations or export regulation – emerge.

For example, the US Department of State controls electro-optical imagery above 0.5 m clear

aperture at 30-m ground sampling distance, synthetic aperture radar, and hyperspectral imagery.

Between 0.35 m and 0.5 m aperture, US companies can seek a licence from the US Department of

Commerce to export imagery. Each operator has to be approved by the National Oceanic and

Atmospheric Administration (NOAA) under the US Department of Commerce. DigitalGlobe did this

in 2014 as it prepared to launch WorldView-3, a satellite that can provide 0.25-m resolution, red-,

green-, blue-wavelength data in some cases, in addition to its near infrared (NIR) and shortwave

infrared (SWIR) bands.

Even more limiting, the Kyl-Bingaman Amendment restricts US collection of imagery over Israel and

NOAA requires a plan specifically for how satellite operators will handle such imagery. In practice,

collection of commercial satellite imagery is highly restricted compared with every other territory on

Earth.

However, consumers unable to buy satellite imagery or information from US companies can

increasingly seek data from other countries, and a future legal challenge will be to keep pace with

rapidly evolving technology in a globalised commercial world.

Another disadvantage of OSINT is that there are fewer experts and practitioners who have been

formally trained in intelligence tradecraft in the open-source community. Intelligence methodologies

stem from curricula formed and honed over decades, and expertise is forged through time spent in

military and government intelligence communities. The disruptive nature of OSINT means that

practitioners may have less experience with tradecraft tools, less rigorous oversight, and no access

to intelligence-grade capabilities such as HUMINT, SIGINT, or military-grade sensors.


On the web Stratollites set to provide persistent-image capability Declassified intelligence offers analytical insight Legal action unlikely to deter militant groups from destroying historic sites






of 7

Author Allison Puccioni is an imagery analyst and founder of the imagery consultancy Armillary Services, LLC. Melissa Hanham is a senior research associate at the James Martin Center for Nonproliferation Studies and META Lab at the Middlebury Institute.

For the full version and more content:

For advertising solutions visit Jane’s Advertising

Jane's Military & Security Assessments Intelligence Centre

This analysis is taken from Jane’s Military & Security Assessments Intelligence Centre, which

delivers comprehensive and reliable country risk and military capabilities information, analysis and

daily insight.

IHS country risk and military capabilities news and analysis is also available within Jane’s

Intelligence Review. To learn more and to subscribe to Jane’s Intelligence Review online, offline

or print visit: http://magazines.ihs.com/

https://ihs.uberflip.com/aerospace-defence-security

http://www.ihs.com/jmsa

http://magazines.ihs.com/

osint transparency raises ethical questions transparency raises ethical questions [content preview...

Documents