Mapping LAWS

Blog 09: Military perspectives on robotics and autonomous systems, Part 1

Mon, 28 Nov 2022 00:00:00 +0000

This post presents a starting point for research we have underway on Australian military perspectives on robotics and autonomous systems. Preliminary findings of this research were presented by Sian at the European International Studies Association conference in September 2022. Through this research we aim to explore the military thinking that is emerging alongside these emerging technologies. We suggest that how militaries imagine the role of these technologies will shape how these technologies develop, how they will be used, and how the armed forces will change. To study military perspectives on robotics and autonomous systems previous research has surveyed defence personnel or analysed planning documents. In our research, we complement and extend current understanding of military perspectives by examining a new data source: a corpus of informal texts written for (and typically by) defence personnel in four Australian defence blogs.

We have gathered blog posts from The Cove, The Forge, Grounded Curiosity, and Williams Foundation Central Blue Blog (see Table 1). These blogs cover a range of issues from life in the military to strategy and tactics. Together, these sources provide a unique and rich picture of Australian military thinking over the past few years and an opportunity to observe the military thinking through the implications of robotics and autonomous systems.

Table 1: Corpus composition

Source	Time period	Word or Punctuation Tokens	Documents
The Cove	2016-2022	1.68 million	1,384
The Forge	2004-2022	0.77 million	324
Grounded Curiosity	2014-2022	0.46 million	262
Central Blue	2012-2022	0.57 million	391

In this first post in our series we tackle a basic question: how prevalent is discussion of RAS compared to other emerging technologies. To do this our analysis quantifies mentions of terms related to robotics and autonomous systems (e.g. RAS, robots, autonomy) and compares their salience to terms associated with other new and emerging technologies. For example, how does robotics and autonomous systems compare to ‘cyber warfare’? Our list of terms (see Table 2) has been compiled from lists of government priority defence-related technologies.

What is striking from comparing these terms is the prominence of RAS-related terms in comparison to other technologies, indicating significant attention to RAS in contemporary Australian military thinking. Word forms of autonomy and robots are very frequent, including the common phrases autonomous systems, autonomous weapons and robotics and autonomous systems. There are a number of technologies related to RAS that are also frequent, including the most frequent technology-related terms Artificial Intelligence and AI and various manifestations of robotic weapons (drones, UAV, unmanned). These early findings indicate a military thinking through the implications of robotics, autonomy and wider questions about the application of AI. We will be picking this up in future posts, which identify when these terms have been used and features of discourse about RAS.

— Sian and Geoff

Table 2: Technology terms and frequency per million tokens

Word	Freq per million tokens	Common phrases
Artificial intelligence / AI	431.1
autonom*	394.6	autonomous systems, autonomous weapons
Cyber	363.3	Cyber security, cyber warfare
Drone	329.9
Robot*	167.8	Robotics and autonomous systems
Satellite*	160.4	Satellite imagery
UAS	116.7
Unmanned	111.5
UAV*	94.6
AWS / LAWS	58.9	AWS and AI
RAS	54.3
Electronic war*	43.7
Machine learning	33.3
Hypersonic*	31.3	Hypersonic weapons
Quantum	22.4	Quantum computing
UGV*	13.2
Nano*	8.9
Internet of things / IoT	4.9
Biotechnology	3.4
Additive manufacturing	3.2

Cite:

Troath, Sian & Ford, Geoff. (2022, November 29). Military perspectives on robotics and autonomous systems, Part 1. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/au-military-perspectives

Blog 08: Drawing autonomy in AUKUS defence documents

Tue, 22 Nov 2022 00:00:00 +0000

I’m moving on from focusing primarily on Australia in previous blogs, to now starting to compare Australia, the United Kingdom, and the United States. As the three Anglophone countries investing significantly in robotics and autonomous systems, and the three AUKUS countries trying to deepen interoperability – I feel they make for a fascinating comparison. This blog will be the first in a series where I compare how autonomy is being discussed in official defence documents across the three states.

This first blog is a very simple one – but nevertheless a good place to start. Here I am simply tracking mentions of autonomy-related words over time. The mentions are taken from all of the official defence concepts or doctrines I could find on Australian, US, and UK defence websites that had at least one relevant mention of autonomy/autonomous/autonomously (ie, not autonomous sanctions). All up, this comprised 63 documents – 13 Australian, 16 British, and 34 American.

On the one hand, the results may reflect most heavily the years in which particular states happened to release official defence documents discussing autonomy (hence a big surge in one year from one state, and little to nothing the next). On the other hand, there is a reasonably noticeable increase over time. Two things I see which jumped out to me – 1) the growing prevalence of Australia over time, and 2) the greater prevalence in Australia than the UK.

In the next blog I’ll be looking in more detail at some key features of these documents, including a comparison of air, ground, and sea. For now, I’ll leave you to explore the visualisation, and the list of sources used below ordered by total mentions of autonomy-related words.

— Sian

Document	Total
14. 2019 Air Force Autonomous Horizons: The Way Forward	1005
7. 2016 US Marine Corps MCWP 3-20.5: Unmanned Aircraft Systems Operations	581
9. 2017 Unmanned Systems Integrated Roadmap 2017-2042	479
37. 2021 Department of the Navy, Unmanned Campaign Framework	388
3. 2015 Autonomous Horizons: System Autonomy in the Air Force - A Path to the Future	349
67. 2020 ADF, Concept for Robotic and Autonomous Systems	275
36. 2021 US Department of Defense, Counter-Small Unmanned Aircraft Systems	263
44. 2017 Joint Doctrine Publication 0-32: Unmanned Aircraft Systems	226
69. 2022 Army Robotic & Autonomous Systems Strategy v.2.0	225
45. 2018 Joint Concept Note 1/18: Human-Machine Teaming	139
65. 2020 Department of Defence, A Method for Ethical AI in Defence	114
32. 2021 Army Futures Command Concept for Protection 2028	101
22. 2020 Army Futures Command Concept for Brigade Combat Team Cross-Domain Maneuver 2028	90
8. 2016 Department of the Navy, Naval Aviation Vision	89
48. 2020 British Army Approach to Robotics and Autonomous Systems: Generating Human-Machine Teams	89
70. 2022 Navy RAS-AI Campaign Plan 2025: Warfare Innovation Navy	89
66. 2020 Royal Australian Navy, RAS-AI Strategy 2040: Warfare Innovation Navy	85
61. 2018 Army Robotic & Autonomous Systems Strategy	81
10. 2018 US Marine Corps MCTP 3-20G: Air Reconnaissance	80
25. 2020 Army Futures Command Concept for Special Operations 2028	69
31. 2021 Army Futures Command Concept for Fires 2028	63
55. 2022 Joint Doctrine Publication 0-30: UK Air Power	58
23. 2020 Army Futures Command Concept for Intelligence 2028	51
47. 2019 Ministry of Defence, Defence Technology Framework	47
1. 2015 Air Force Future Operating Concept: A View of the Air Force in 2035	44
29. 2021 Army Futures Command Concept for Command and Control 2028: Pursuing Decision Dominance	39
54. 2021 Joint Concept Note 3/21: Future Joint Personnel Recovery	36
52. 2021 Ministry of Defence CP 411, Defence in a Competitive Age	31
38. 2021 Navy Aviation Vision 2030-2035	29
39. 2017 Ministry of Defence Joint Concept Note 1/17: Future Force Concept	28
30. 2021 Army Futures Command Concept for Cyberspace and Electromagnetic Operations 2028	27
56. 2022 Royal Navy, Maritime Operating Concept MarOpC: The Maritime Force Contribution to the Integrated Operating Concept	27
6. 2016 Joint Operating Environment JOE 2035: The Joint Force in a Contested and Disordered World	26
42. 2017 Joint Doctrine Publication 0-30: UK Air and Space Power	26
11. 2018 Army the U.S. Army in Multi-Domain Operations 2028	23
26. 2020 United States Marine Corps, Department of the Navy, and Coast Guard, Advantage at Sea: Prevailing with Integrated All-Domain Naval Power	23
12. 2018 U.S. Army Concept: Multi-Domain Combined Arms Operations at Echelons Above Brigade 2025-2045	22
24. 2020 Army Futures Command Concept for Maneuver in Multi-Domain Operations 2028	21
58. 2016 Department of Defence, 2016 Defence White Paper	16
4. 2016 Air Force Doctrine Publication 3-0: Operations and Planning	14
21. 2020 Air Force Doctrine Publication 3-59: Weather Operations	14
15. 2019 Air Force Doctrine Publication 3-01: Counterair Operations	13
50. 2020 Joint Doctrine Note 1/20: Air Manoeuvre	13
63. 2020 Department of Defence, 2020 Force Structure Plan	13
19. 2020 Air Force Doctrine Publication 3-03: Counterland Operations	12
33. 2021 Army Modernization Strategy: Investing in the Future	12
28. 2021 Air Force Doctrine Publication 3-52: Airspace Control	11
57. 2015 Department of Defence, Guarding Against Uncertainty: Australian Attitudes to Defence	11
60. 2018 ADF Concept for Command and Control of the Future Force	11
34. 2021 Army Chief of Staff Paper #1, Army Multi-Domain Transformation: Ready to Win in Competition and Conflict	10
49. 2020 Joint Concept Note 1/20: Multi-Domain Integration	10
18. 2019 Army Modernization Strategy: Investing in the Future	9
40. 2017 Joint Concept Note 2/17: Future of Command and Control	8
5. 2016 United States Joint Staff Joint Force Development (J7) – Future Joint Force Development: Cross-Domain Synergy in Joint Operations	7
2. 2015 Air Force Doctrine Publication 2-0: Globally Integrated ISR Operations	6
53. 2021 Ministry of Defence, Integrated Operating Concept	6
62. 2019 Royal Australian Air Force, At the Edge: Exploring and Exploiting our Fifth-Generation Edges	6
17. 2019 Air Force Doctrine Publication 3-10: Force Protection	5
51. 2021 Defence Experimentation for Force Development Handbook	5
59. 2017 Royal Australian Navy, Mercator: Maritime Domain Strategy 2040	5
13. 2018 Summary of the 2018 National Defense Strategy of the United States of America: Sharpening the American Military’s Competitive Edge	4
27. 2021 Air Force Doctrine Publication Note 1-21: Agile Combat Employment	4
35. 2021 The Army Unified Network Plan 2021: The Army Unified Network Plan Enabling Multi-Domain Operations	4
41. 2017 Joint Doctrine Note 1/17: Joint Theatre Entry	4
46. 2018 Joint Doctrine Note 1/18: Cyber and Electromagnetic Activities	4
20. 2020 Air Force Doctrine Publication 3-05: Special Operations	3
64. 2020 Department of Defence, 2020 Defence Strategic Update	3
16. 2019 Air Force Doctrine Publication 3-04: Countersea Operations	2
43. 2017 Joint Doctrine Publication 0-20: UK Land Power	2
68. 2021 ADF Philosophical Doctrine: Military Ethics	1

Cite:

Troath, Sian. (2022, November 22). Drawing autonomy in AUKUS defence documents. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/drawing-autonomy-aukus

Blog 07: RAS, the budget, and the election

Tue, 10 May 2022 00:00:00 +0000

Photo: Minister for Defence Industry Melissa Price at Orbital UAV

Defence matters have been surprisingly and unusually prevalent during the Australian federal election campaign these past few weeks. Both major parties have seized on the announcement of a military cooperation agreement between China and the Solomon Islands as an opportunity to turn to their electoral advantage. Amidst this has been a number of defence spending announcements, following some significant budget announcements prior to the start of the campaign. It seems a good time, then, to take stock of what has been happening with regards to robotics and autonomous systems (RAS) during and since the budget, and the key potential implications of the election outcome.

The big headline of the budget in defence terms was the announcement of REDSPICE (resilience, effects, defence, space, intelligence, cyber, enablers), the single most significant investment in the Australian Signals Directorate (ASD). Autonomous systems feature only as one of the challenges that REDSPICE is responding to, with the funding itself focused more on cyber matters. Indeed, the allocation of funds to REDSPICE was, in part, a result of the cancellation of the AIR 7003 Phase 1 project to deliver the MQ-9B SkyGuardian from General Atomics, an armed medium altitude long endurance remotely piloted aircraft system – a controversial decision critiqued by analysts and defence industry. At the same time, exploration of unmanned aerial capabilities for the civil maritime environment was also paused in preference for continuing manned flights. Home Affairs Secretary Pezzullo stated ‘that the state of unmanned aerial capability globally and in use by defence was not going to meet our requirements … in a civil maritime environment’.

While some RAS projects seem to be struggling, others are doing better. For instance, a suite of funding announcements from the Minister for Defence Industry on the 6th of April included several contracts focusing on the development of uncrewed systems. The golden child of Australia in autonomous systems, however, is the Ghost Bat, formerly known as the Loyal Wingman program. It is an unmanned aircraft, designed to fly alongside manned and unmanned aircraft. The Ghost Bat was recently praised by the US Air Force Secretary Frank Kendall, who stated that ‘we’re going to use technologies that are coming out of programs like the Australian Loyal Wingman program and others, and we’re going to integrate those into our operational capability’.

Given this sea of hits and misses, what differences might we expect the election to bring?

The major parties are in broad agreement about much of defence policy, including a commitment to higher defence spending and a more or less shared perspective on Australia’s deteriorating security environment. Both are committed to the US alliance and to the AUKUS partnership. The usual differences remains, for instance with Labor also more interesting in boosting diplomatic capacity.

Both come to autonomous systems from the same point of view of seeking military advantage in an environment of strategic instability, managing alliance relations, and trying to plug capability gaps in the near term.

In terms of differences, Labor has stated they would consider reinstating the funding for the SkyGuardian program to acquire up 12 armed drones. Shadow Defence Minister Brendan O’Connor said ‘we’d be examining that as a matter of urgency if elected’, and argued uncrewed armed drones were an area where Australia is ‘deficient’.

Labor has also announced they will create the Advanced Strategic Research Agency (ASRA), the Australian answer to DARPA. It will ‘fund pivotal research in breakthrough technologies for national security’, with an aim to ‘boost Australia’s involvement in technology sharing and research and development, through the new AUKUS partnership’. This could provide a vehicle for boosting research and development of RAS, but it will depend on priorities.

These policies compare with the Coalition’s focus on cyber under the auspices of the ASD, their existing Action Plan for Critical Technologies, and continuing to develop the AUKUS partnership.

Aside from these details, there isn’t a huge amount of difference evident between the two parties at this stage. Both are, in principal, committed to expanding defence spending and focusing on new and emerging technologies, particularly in the context of the AUKUS partnership. The biggest difference in how autonomous systems develop, I believe, will lie much more in the practicalities of how these shared principles will be enacted. In particular, research and development and higher education policies will be influential, as they will shape the research that underpins the relevant technologies. Another significant factor is external changes – whether this be changes to the security environment that encourage higher defence spending or different acquisition priorities, or situations such as the use of armed drones in Ukraine prompting more attention on RAS.

In short, the election does little more than emphasise how little daylight there is between the two major parties on defence policy. It demonstrates we will see a doubling down on AUKUS, the US alliance, and a continuing focus on new and emerging technologies with national security implications. The result of the election may make a difference in the machinations of how such things are accomplished, but little difference to what is trying to be accomplished.

— Sian

Cite:

Troath, Sian. (2022, May 10). RAS, the budget, and the election. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/ras-and-the-election

Blog 06: Drawing Autonomy in Australian Defence Documents II

Wed, 16 Feb 2022 00:00:00 +0000

In my first blog in the series on drawing autonomy in Australian defence documents I focused on the themes and frequency differences between the services when it came to their use of the word autonomous in key planning documents.

This time, I’ve taken a broader sweep of documents with an aim of focusing on theme and frequency in general, rather than a service-based comparison. This enables broader insights into the place of RAS in defence thinking – as we can also see the emphasis placed on RAS in broader planning documents, and the presence or absence of the themes identified in the previous blog.

This time, the documents used include:

2016, Department of Defence, Defence Industry Policy Statement

2016, Department of Defence, Integrated Investment Program

2018, ADF, ADF Concept for Command and Control of the Future Force

2018, Australian Army, Robotic and Autonomous Systems Strategy

2019, RAAF, At the Edge: Exploring and Exploiting our Fifth Generation Edges

2020, ADF, Concept for Robotic and Autonomous Systems

2020, RAN, RAS-AI Strategy 2040: Warfare Innovation Navy

2020, Department of Defence, Force Structure Plan

2020, Department of Defence, Defence Strategic Update

2020, Kate Devitt, Michael Gan, Jason Scholz, and Robert Bolia, A Method for Ethical AI in Defence

I have used the same themes as before, as you’ll see in the key – although including the two which weren’t present in the first four documents used the previous blog. This time, I have also categorised the documents into four groupings: those relating to defence industry (1, 2), defence planning documents which are broad in scope(3, 8, 9), defence planning documents that are more specifically focused on RAS (4, 5, 6, 7), and documents relating to ethics (10).

A quick final note on process before I leave you to explore the data viz – each document was searched for the word autonomous, with exclusions being made for being in the name of research centres, for being in the reference list, and for being in appendices. I did put some thought into the symbols being thematic – the money sign and tick are obvious, but the circle for broad defence planning is inspired by the OODA loop, and the square for RAS defence planning is a very simplified classic robot head.

Rather than say any more about what this process revealed to me, I’d rather leave you to explore the data yourself – in the spirit of Lupi and Posavec’s approach to data humanism. Please do let me know if you have any thoughts on the choices of themes and documents, or if there’s any other angle you would like to see explored in future blogs.

— Sian

Cite:

Troath, Sian. (2022, February 16). Drawing Autonomy in Australian Defence Documents II. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/drawing-autonomy-ii

Blog 05: New Zealand’s autonomous weapons challenge

Fri, 24 Dec 2021 00:00:00 +0000

As 2021 draws to a close another meeting of the Convention on Conventional Weapons (CCW) ends without making any meaningful progress on the issue of lethal autonomous weapons systems (or LAWS). Much has already been said about the general sense of inertia at the CCW meetings, so I want to focus specifically on the challenges faced by Aotearoa New Zealand in its attempt to play a more assertive role in these debates.

Last week I was asked to make some comments on a story being produced by TVNZ journalist Logan Church on New Zealand’s stance on ‘killer robots’. Earlier in the week Church had run an initial story on the active stance being taken by New Zealand’s Minister for Arms Control and Disarmament, Phil Twyford, in favour of a regulation/ban framework for autonomous weapons under international law.

For his follow-up piece, Church interviewed prominent computer scientist and AI expert Stuart Russell, who has spoken out against the development of ‘killer robots’ for many years. Russell opined that New Zealand’s position within the Five Eyes intelligence network (alongside the US, UK, Canada and Australia) could make it an influential voice in challenging the AI/autonomy-focused weapons programmes in the partner countries. For my part, I argued the opposite: that if anything, New Zealand’s place within Five Eyes will make it difficult for this country to make and sustain a commitment to not using LAWS. There are a number of reasons for this.

From a political perspective, it looks painfully naïve to think that the United States, or even Australia, would take cues from New Zealand on the regulation of their military industries and the ethics of their military deployments. New Zealand has been a target of some resentment and scorn from their Five Eyes partners over recent years for not being tough enough on China, and fragments of antipathy towards New Zealand’s strong history of anti-nuclear activism, which ended New Zealand’s formal alliance with the US, still exist in the halls of power in Washington. Whether there is appetite to potentially create another defence rift between the countries over these new technologies remains to be seen.

The possibilities of such a rift are real. The US, UK, and Australia are all investing very heavily in AI-powered weapons technologies, many of which have autonomous capabilities, in the belief that this is the best way to find an edge (or ‘offset’) if the developing tensions with China spill over into military conflict. There is a very strong expectation that New Zealand will stand shoulder-to-shoulder with its ‘traditional’ defence partners against Chinese aggression in the Indo-Pacific, an expectation that New Zealand’s most recent Defence Assessment seems to clearly affirm. It is for this reason that Twyford’s Cabinet paper on LAWS allows for ‘maintaining interoperability with our key defence partners’ and acknowledges that ‘our policy settings will need to take into account the prospects of operating alongside, and being able to communicate meaningfully with, AWS deployed militaries.’ This looks like a very significant carve-out designed to placate those key defence partners, and one which potentially undermines the whole commitment to bans or regulation of LAWS.

But the potential pitfalls may run even deeper than this. As a Five Eyes partner, New Zealand is responsible for gathering and sharing digital communications and intercepts from the South Pacific region and feeding them into US-based analysis systems. There is no question that this mass of data, along with that collected by other partners, could be used to train machine learning systems deployed by the US or Australia and could potentially be used as inputs into algorithms for targeting or attack decisions in war. We have already seen how metadata has been deployed by the US to identify targets and conduct ‘signature strikes’ with drones in the war on terror and this is only likely to increase with the development of autonomous technologies.

In a future scenario in which New Zealand has signed and ratified a ban on LAWS and then conducts military operations alongside LAWS-equipped militaries, or continues to supply intelligence data that the weapons systems need to function, accusations of a breach of their legal commitments or at the very least hypocrisy would be hard to deny. It might be argued that so long as the New Zealand military doesn’t kill people with its own autonomous technologies, then the spirit of the commitment would be maintained. But the lines here are blurry, as they are in all matters related to these weapons technologies and algorithmic war. It is a relatively cost-free exercise to recognise and call out the moral and ethical problems of LAWS, but for New Zealand to consistently adhere to those commitments will require a more radical revision of ‘traditional’ security settings.

When taking its anti-nuclear stance in the 1980s, the New Zealand Government was prepared to hold a strong and consistent line, right down to the denial of the US ship visit which led to the breakdown of ANZUS. There was a heavy price to be paid for that stance and the Government was willing to pay it. Would today’s New Zealand Government stand behind its commitments to ban LAWS to the extent that it meant withdrawing from the Five Eyes intelligence arrangement, or lead to another intense souring of relations with the US and – this time – Australia? It’s very hard to see such a situation eventuating in the current political context, which will make the playing out of Twyford’s anti-LAWS agenda very interesting to watch over 2022 and the years beyond.

— Jeremy

Cite:

Moses, Jeremy. (2021, December 24). New Zealand’s autonomous weapons challenge. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/nz-challenge

Blog 04: Drawing Autonomy in Australian Defence Documents

Thu, 14 Oct 2021 00:00:00 +0000

From 2018 to 2020 the Australian Army, Air Force (RAAF), Navy (RAN), and the Australian Defence Force (ADF) as a whole released either specific robotic and/or autonomy concepts, or guidance documents which heavily involve autonomous systems (the latter in the case of the RAAF):

2018: Australian Army, Robotic and Autonomous Systems Strategy

2019: RAAF, At the Edge: Exploring and Exploiting our Fifth Generation Edges

2020: ADF, Concept for Robotic and Autonomous Systems

2020: RAN, RAS-AI Strategy 2040: Warfare Innovation Navy

This suite of documents is interesting, revealing insights into how the services are approaching new and emerging technologies. They have appeared in the context of increasing funding being allocated to ‘trusted autonomous systems’, and various defence planning documents such as the 2016 Defence Industry Policy Statement identifying trusted autonomous systems as a priority area.

This blog will be the first in a series using data visualisation techniques drawn from the work of Giorgia Lupi and Stefanie Posavec to explore autonomy in Australian defence documents. Later work will explore different themes, as well as use a broader collection of documents outside those which have a specific focus on robotics, AI, and autonomous systems.

But first let’s start with the basics: mentions of the word autonomous in these four documents.

I have looked at the uses of the word ‘autonomous’ in these documents, including titles and in references. The only exception was from the ADF concept: it has a header on every single page with the word autonomous which, at 74 pages long, rather skewed the data. I have also not included variations of autonomous (such as autonomy or autonomously) at this point.

I was interested in the context in which the word autonomous was mentioned, and so I developed a number of categories I expected might come up: threat, opportunity, challenge, priority, neutral/descriptive, regarding aquisition, regarding ethical and legal challenges, regarding regulation, and regarding technical features/definitional questions. Due to no hits, ‘priority’ and ‘regarding aquisition’ were dropped. I expect these will make a return in future blogs which use an expanded selection of documents.

The categorisation process was also somewhat subjective – as the process of categorisation always is. This is in fitting with the spirit of Lupi’s ‘data humanism’, which emphasises that data visualisations should tell stories rather than seek to represent objective truths. Where mentions of the word autonomous crossed multiple categories, I picked the one which seemed to me to be the strongest theme. In one instance, a sentence in which robotics and autonomous systems are described as ‘both a threat and an opportunity for Defence’ has been double categorised as ‘threat’ and ‘opportunity.’

Below you will find a key for understanding the data visualisation and the data visualisation itself. I aim to give a sense of the differences between the different documents, both in quantity and category of mentions of the word autonomous.

— Sian

Cite:

Troath, Sian. (2021, October 14). Drawing Autonomy in Australian Defence Documents. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/drawing-autonomy

Blog 03: Explore representations of robots in popular culture with Robot Dreams

Wed, 22 Sep 2021 00:00:00 +0000

Kia ora koutou,

References to the killer robots of popular culture are a recurring feature in public discussion of robots and autonomous weapons. In our recent research we’ve observed how Twitter users play on killer robot references, including mentions of Terminator and Black Mirror, when referring to, Spot, Boston Dynamics’ robot dog. In public discussion of autonomous weapons these references occur in texts intended for general audiences, for example news media reports and popular non-fiction books on autonomous weapons, but also in academic, government and activist writing. The coalition of activist groups advocating for regulation of autonomous weapons even appeals to public familiarity with killer robots in their name: the Campaign to Stop Killer Robots.

The Mapping LAWS project is mapping and analysing features of debates about autonomous weapons, including these popular culture references. Robot Dreams is an interface to explore representations of robots in cinematic imagination and draw attention to themes related to war, violence and threat.

Robot Dreams plays a sequence of sentences mentioning robots from a data-set of over 82,000 movie plots created from Wikipedia. As the robotic voices speak each sentence a visualisation of possible related themes appears. It takes over four hours to watch and listen to more than 1800 robot references. This is academic research, but it is intended as an engaging and playful way to think about robots on film.

This post is an introduction to Robot Dreams and some of the ideas behind it. If you haven’t already, please go take a look at Robot Dreams.

Ideas

When thinking about how to use data to communicate with other researchers and the public, the Mapping LAWS project is drawing inspiration from the work of information designer Giorgia Lupi. Lupi’s advocacy for data humanism challenges the standard visual language of data presentation that prioritises simplifying representations of complex phenomena, but which also reduces possibilities for deep engagement and thinking. The design-driven approach she proposes aims to introduce audiences to the complexity of data, to depict relevant context, and open data up for audiences to explore.

Robot Dreams makes use of the ability to animate graphics and play audio via a web interface to look and listen and think about representations of robots on film. By listening to the cold robotic voices reading movie plot excerpts in a sequence we have the opportunity to experience the prevalence of themes in a unique way, to inspect the range of meanings assigned by the machine-assigned themes and explore connections across movies. The visualisation allows us to see the relative prominence of themes related to robots depicted as circles - larger circles indicate more pervasive themes.

Rather than obscuring the imperfections of the data through simplifying quantifications and abstracted views, Lupi suggests:

“We should learn how to include and render the more qualitative and nuanced aspects of data. We should experiment with how to visualize uncertainty, possible errors and imperfections in our data”

Something interesting and unexpected about Robot Dreams is the ability to engage with the data in this qualitative way.

It is relatively common with computer-assisted text analysis to focus on quantifications: in this case, we could have classified sentences to determine the proportion that indicate war and violence, reporting this with an error rate to abstract the imperfections of the data. The unsupervised, exploratory approach I’ve used highlights a range of possible themes. These themes have been extracted using Empath and its default category dictionaries. The audio presentation of each excerpt with a visual representation of associated themes allows us to evaluate the themes assigned, to engage with the nitty gritty of the data, and perceive the range of meanings associated with words used in the excerpts.

While this method certainly has its limitations, it is interesting in that it opens up the audience to the range of possible meanings, allowing the very human reaction of thinking about and arguing with the associated themes. The randomness and the sequential presentation of the excerpts also allows new views of the data - to notice new features and make new connections.

Madness

There is a serious point to Robot Dreams - to explore themes related to robots in popular culture and depict the prevalence of war, violence and threat - however, listening to the random sequence of excerpts is fun and perhaps a little silly. That is ok!

So go take a look at Robot Dreams and share it around - I’m interested in your reaction. Email me your thoughts!

— Geoff

Method

Robot Dreams has been built with a bit of Jupyter Notebook data processing and hand-crafting of the web interface. Here is a brief rundown of how this was built:

Using Wikipedia’s API, movie articles were collated by querying for pages featuring the film infobox. Then 82,000 movie page sections with titles like “Plot” or “Synopsis” were downloaded.

These plots were processed with Spacy to segment sentences and any sentence featuring robot* were extracted. This doesn’t include related words (e.g. cyborg, android, mech or drone) or robots with names (e.g. Terminator).

Then Empath was run across the sentences to extract themes. Words may (and often do) trigger multiple themes. Empath’s themes are based on a language model that embeds semantic relationships between words. This language model has been trained by the Empath authors on a data-set of fiction plots, so there is some alignment of the language domain (i.e. fiction and movie plots). To avoid over-emphasising themes cued by robot, robots etc, these words were removed.

The pyttsx3 library was used to convert text to robot speech. The robot voices used were part of the accessibility features of the Windows operating system. While there are more realistic machine-generated voices available, I was interested in voices that sounded machine-like.

Hierarchical clustering was used to group related Empath themes and this influenced the position of the circles. For example, war, fight, kill, and military are semantically related and are grouped in the visualisation. I used scikit-learn for the theme clustering. Using a word2vec model via Gensim, vector representations of words within each theme were averaged and then ‘close’ themes grouped.

The dream - the collection of circles indicating the prevalence of different themes - is a packed circle chart. Plotly, which allows lots of control of output and layout, doesn’t produce packed circle charts. However, the circlify library allows calculating a packed circle chart layout. I’ve shared a notebook to show how packed circle charts can be drawn with Plotly.

The web interface was hand-crafted (with love), using SVG output from Plotly, a robot head image and the animation library anime.js.

Acknowledgement and Attribution:

Thanks to Nicki Cartlidge, a soon-to-be PhD candidate at the University of Canterbury. A conversation with her about her planned PhD project on sonification made me think in new ways about representing the data.

The image of the robot is based on an file from Noun Project by Designfour. I’ve adapted this to make it work for the design and animation.

Cite:

Ford, Geoffrey. (2021, September 22). Explore representations of robots in popular culture with Robot Dreams. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/blog-robot-dreams

Blog 02: Human Starts the Loop?

Fri, 27 Aug 2021 00:00:00 +0000

Kia ora koutou,

For my part in the Marsden project one of the things I will be looking at is Australia as a case study. I will be exploring what debates are being had and by who related to lethal autonomous weapons (LAWS). I have only recently started on the project, and so my research is very preliminary at this stage.

In beginning to research the LAWS space in Australia, however, there is an interesting tidbit which has stood out that I’d like to highlight.

There is a term which appears in Australian planning documents, which I had not seen in the US or activist literature: human starts the loop. This sent me on a divergent little hunt to explore whether or not this term had been used by anyone else.

For those not familiar with the literature, questions surrounding to what degree a human should be involved in the decision-making of LAWS has played a central role in military, political, and activist debates. A number of parties have settled on the concept of meaningful human control as central to their arguments around banning or regulating these technologies. A definition for meaningful human control has yet to be agreed upon.

This debate on meaningful human control is related to the concepts of human in the loop, human on the loop, and human out of the loop – all of which are common ways to talk about the degree of autonomy a particular technology has, and the relationship between people and that technology. In the Australian Defence Force’s Concept for Robotics and Autonomous Systems published in 2020, human in the loop is defined as ‘the system performs some functions independently but requires a human to perform functions that complete the system’s task cycle’, while human on the loop is defined as ‘the system performs all functions autonomously but a human may intervene to stop or modify the outcome before the task is complete’. Human out of the loop is not used in the same document, but as is evident in the term it implies some form of full autonomy.

Human starts the loop is, then, an intriguing term to find popping up in Australian Defence planning documents. I have found eight mentions of the term so far, ranging from June 2018 to November 2020. There are two defence blogs where the term is mentioned off-hand, three pieces by Senior Research Fellow at the Institute for National Strategic Studies, T.X. Hammes, and three Australian defence documents where the term is used. Using data humanism, I have depicted the emergence of the term human starts the loop below:

The first detailed mention is the ADF Concept for Command and Control of the Future Force. In this document, they categorise types of human-machine teaming as full human control, human in the loop, human on the loop, and human starts the loop. Human starts the loop is defined as ‘a human sets operational parameters and initiates the machines operation; the machine requires no further human interaction to complete its task’.

This is the same categorisation used in the later 2020 Concept for Robotic and Autonomous Systems. Going into greater detail, they match this ‘control context’ with the ‘technical context’ to define four categories of robotics and autonomous systems: remote systems, automatic systems, autonomic systems, and autonomous systems. Autonomous systems are in the technical context ‘goal based reasoning’ and in the control context slide across from the latter half of human on the loop and through to human starts the loop. In listing examples of each kind of system, AlphaZero is the only example of an autonomous system. The Roomba automatic vacuum and Tesla autopilot are listed as examples of autonomic, as while they fall under the umbrella of human starts the loop in the control context, in the technical context they fall under task based reasoning, rather than the goal based reasoning required for a system to be considered autonomous. The Navy’s RAS-AI Strategy 2040 merely lists a definition of human starts the loop in its index, it does not use the term in the body of the document.

T.X. Hammes argues that human out of the loop is ‘a bad definition’, given that humans design and build autonomous systems, providing the parameters by which they will operate. Therefore, he claims, human starts the loop is a more useful term, accurately reflecting the relationship between people and autonomous systems.

Where this all becomes interesting is when it is placed in the context of the activist language around meaningful human control. I can foresee an argument being made for meaningful human control existing in the technical context, while brushing the control context aside with the use of the term human starts the loop. This is not to say such an argument would necessarily be successfully convincing, but rather that this is where the focus may be placed, likely taking place under the auspices of trusted autonomy. ‘Trusted autonomous systems’ are a focus area for Defence. Work in this space examines how to make sure defence personnel trust the autonomous systems they will be using, and improve public perceptions of autonomy. This includes making the case that those who operate the technology both trust that technology and understand what it is doing, and that the technology has been designed and will be used in accordance with international law.

As part of my research I will be keeping an eye on the use of ‘human starts the loop’ and its relationship with the strong focus in Australia on trusted autonomous systems. Stay tuned for any developments!

— Sian

Image Credit: The diagram above is from the ADF’s 2020 document: Concept for Robotic and Autonomous Systems

Cite:

Troath, Sian. (2021, August 27). Human Starts the Loop? Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/human-starts-the-loop

Blog 01: Introducing the Mapping Laws Project

Thu, 05 Aug 2021 00:00:00 +0000

Kia ora koutou and welcome to our blog!

We are a group of researchers based at the University of Canterbury (Jeremy Moses, Geoffrey Ford, and Sian Troath) and Eastern Kentucky University (Amy Fletcher) working on a research project that digs into the debate about lethal autonomous weapons systems (LAWS) or ‘killer robots’.

Our project, which has been funded for a 3 year period by New Zealand’s Marsden Fund, aims to develop a detailed ‘issue map’ of the debate on killer robots. Our first task has been to collect a huge range of literature on the subject from the media, academia, the military, government, inter-governmental and non-governmental organisations, popular culture, and the weapons industry. We’re now analysing that material in order to understand the key themes that are being discussed in each area and the differences and connections between them.

This data will be presented in formats designed to engage a range of audiences that are interested in this topic. In developing our data vizualizations we are taking inspiration from the work of Giorgia Lupi and her concept of ‘data humanism’, which aims toward drawing the viewer in to a story rather than claiming to represent absolute truths. We plan to present some preliminary visual overviews of the data in the coming months.

In the meantime we will regularly update this blog with snippets of the research we are undertaking, including some smaller-scale, hand-drawn vizualizations of elements of the killer robots debate that we are working on or that may be of particular interest at a given moment.

We’re looking forward to the research journey ahead and hope you will join us for the ride!

— Jeremy, Amy, Geoff, and Sian

Cite:

Moses, Jeremy. (2021, August 5). Introducing the Mapping Laws Project. Mapping LAWS: Issue mapping and analysing the lethal autonomous weapons debate. https://mappinglaws.net/blog-introduction