23 March 2019

Pentagon’s new strategy on artificial intelligence and killer robots’ ethic

Niculae Iancu

Image source: Mediafax

The US Defence Department has announced, in a press conference held on 12th of February, the launch of its first strategy related to artificial intelligence. This announcement comes after a series of premieres confirming artificial intelligence’s important role across Washington’s current vision. Maintaining the American Leadership in artificial intelligence has some conceptual valences, and its military dimension reveals Washington’s concern on the global competition in this field, especially on Russia and China’s increasing interest for investing more in developing their own artificial intelligence capacity.

These states are still at the beginning of this journey of finding and valuing artificial intelligence’s benefits. We will see if at the end of this journey there is that world of killer robots or global wealth and prosperity the human society wishes to find.

The great actors of the competition for artificial intelligence leadership

Pentagon has announced, in the middle of this month, the launch of its first strategy related to artificial intelligence. The announcement was sync with the implementation, a day before, of President Donald Trump’s executive order for the maintenance of American Leadership in artificial intelligence. “The executive order is paramount for our country to remain the leader in AI and will not only increase the prosperity but also enhance our national security”, stated the chief information officer across the public presentation of the strategy, together with Lt.Gen. Jack Shanahan, the first director of the DOD Joint Artificial Intelligence Center of the Department of Defence. The military strategy is also the first US document called “strategy” in artificial intelligence after a series of “preparations” and “plans” adopted starting with 2016.

The multitude of these perspective and action premieres on the role of artificial intelligence in maintaining the global American military leadership highlights, if there was still any doubt, Washington’s concern on the significant changes of the international security environment and, particularly, the concern on Russia and China’s interest in investing more in developing their own artificial intelligence capacities (AI).

Starting with 2017, other 20 states were assuming strategies, strategic plans or budgets dedicated to artificial intelligence. Among these stays China, which communicates better and better on strategic matters, however, Russia is missing, as expected, given the traditional opacity of Kremlin’s elites. On its turn, the European Commission adopted in April 2018 the “Communication on artificial intelligence” and it was establishing a series of measures for the increase of Union’s industrial and technological capacities in AI field.

Hence, it can be said that it was the start for AI leadership in the global race. The first country to adopt a national strategy for artificial intelligence, in March 2017, is Canada. The “pan-Canadian artificial intelligence strategy” has a budget worth of 125 million CAD for the first five years of implementation. UNESCO mentions in a press release, back in November 2018, which were the reasons behind the document adopted by the government in Ottawa. “Canada had the talent advantage but we needed to act quickly to maintain that lead. International demand for talent, especially from the USA, was putting Canada’s prior investments in AI research and talent development at risk. There were concerns in both government and in the private sector that this brain drain would compromise Canada’s capacity to become early adopters of this disruptive new technology.” Hence, the Canadian strategy in the AI field is purely humanistic, and its central benchmarks are research and education.  

China has revealed its ambition level in becoming a global leader in AI research, development and technologies application, in July 2017. “A next generation artificial intelligence development plan” of the Chinese government is a systemic complex vision over the field, from education to security. The plan establishes targets for the decrease of Chinese technological dependency on foreign technology under 30% until 2015, by promoting the “made in China” vision. Basically, the steps to be followed would be “aligning with the competition until 2020”, taking some leader positions in AI technologies until 2025 and gaining the AI leadership until 2030. In order to reach the latter, Beijing wants to redirect the global flows of intelligence migration towards the continental coast of East China Sea and wants to allocate budgets worth of billions of dollars to support AI innovative projects.

All these plans, from aspirations to numbers, should be considered even given Trump’s Administration force measures against the Chinese intellectual espionage inside the American universities and its determination to protect the intellectual property and trade transaction control between Washington and Beijing, as part of a tough competition that many analysts consider to be like a “new Cold War”.

India offers a more balanced vision over its intentions on artificial intelligence. The “National Strategy for artificial intelligence #AIFORALL” refers, also, to Indian investments orientation towards education and research, but it stops at the alignment desideratum with the international developments in the field, following the growth of economic and social benefits of AI technologies. In the competitional logic, India aims at becoming the “garage for 40% of the world”, which means the solution provider of choice for the emerging and developing economies across the globe. The effects should be found in an annual increase of 1,3 % of the Indian economy until 2035.

Japan is the second country to adopt an AI strategy after Canada. In March 2017 it was launched the “Artificial Intelligence Technology Strategy”, which presents maybe the most pragmatic and, implicitly, the largest successful potential, a conception in valorizing AI opportunities. Japanese’s successful model which lead to hijacking the audio-video, auto or communications industries markets in the last decades is also found in AI vision as well, shaped in two-steps: AI research -development and industrialization. Tokyo expects its strategy to help it switch to “5.0 Japanese Society”, a productive ecosystem, healthy and versatile, which interconnects various fields and represents a symbol for Japan’s unique innovation capacity to be permanently a step ahead the global evolutions, if we consider the current concepts, 4.0 industrial revolution, and the 4.0 globalization phenomenon.

Although lacking of a public AI strategy, Russia cannot make an exception at the global competition analyses for AI leadership. Moscow has a central role in the majority of Western evaluations on military applications of artificial intelligence, and the supposed Russian armament systems with high-action and decision autonomy are increasing the concerns regarding a new arms race, more sophisticated and with a more devastating potential than the Star War at the end of the Cold War. President Vladimir Putin was mentioning in the public space, back in September 2017, that "Whoever becomes the leader in this sphere will become the ruler of the world” and that it should be prevented the monopolization of this power in a “pair of hands” because it involves huge opportunities, but also threats that are difficult to foresee today. Furthermore, Putin was stating, with certain altruism, that if Russia will become the leader in the development of artificial intelligence “we will share our technology with the rest of the world like we are doing now with atomic and nuclear technology".

The European Commission was adopting the “Communication on artificial intelligence” in April 2018. The communication mentions that 24 EU member states- Romania is not listed among them- and Norway have “committed to joining forces in AI”. Starting from this “strong political commitment”, EU will make efforts to become “competitive in AI landscape”, says Commission’s communication, for no one to “be left behind in the digital transformation” and to “found new technologies based on European values”. EU priorities are European research-development and SMEs capacity to adapt to and adopt AI technologies. The communication admits that in 2016 EU was way behind regarding the private investments in AI field, with a budget of around 2,4-3,2 billion euro, comparing to 6,5- 9,7 billion in Asia and 12,1-18,6 billion in North America. On the other hand, it is mentioned the continuity of robotics investments starting with 2004, which allowed “Europe to become the global leader in this field” and it is expressed a commitment to increase the budget for projects with AI application, financed by community funds.

Along with the common effort to increase EU’s role in AI, many member states have adopted national strategic documents related to this field, like Denmark, Finland, France, Germany Great Britain, Poland and Sweden. Also, Denmark, Estonia, Finland, Faeroe Islands, Island, Latvia, Lithuania, Norway and Sweden have signed the “Artificial Intelligence in Nordic-Baltic region”, wherein they are praising the “Communication on artificial intelligence” of the European Commission and expressing their support for “AI Cooperation” signed by the 24-member states and Norway, in order to assume a common approach for all AI challenges.

The strategies on artificial intelligence and the importance governments are giving to this field are increasing. AI’s actors have already declared their intentions, the funding has started to follow, way coherently, more and more ambitious objectives, emerged from AI’s technologies benefits, but also from aspiring to a significant place in the global decisional and influential architecture of artificial intelligence future. We should not neglect that, as history has shown us, if there is a great ambition and intention, competition is inevitable and collisions are inherent. How great actors will choose to act inside AI space will depend on the magnitude and violence of these collisions, which will not influence only our daily life, but the purpose of our existence.  

The mainstreams of president Donald Trump’s executive order for the maintenance of American leadership in artificial intelligence

President Trump’s firm measures and unilateral decisions regarding intellectual property protection and American global technological leadership are placing artificial intelligence in the center of its mandate’s national security concerns. Trump’s Administration is shaping the strategy regarding artificial intelligence starting from two essential premises. Firstly, the United States think that they have the global leadership position in research, development and production of artificial intelligence-based technology. Secondly, artificial intelligence “promises” to become the main factor of “United States’ economy growth”, of “national and economic security consolidation” and “life quality improvement”.

The strategic directions before issuing the executive order were adopted by Obama, also in two consecutive days, in October 2016. These were the first documents to have strategic relevance for the American AI ecosystem in artificial intelligence. On 12th of October, it was emerging the “Preparing for the future of artificial intelligence”, and the day after that, 13th , it was adopted the “The national artificial intelligence research and development strategic plan”.

The first document mentions the strategic priority of artificial intelligence for the United States and lists 23 recommendation for the consolidation of the United States as the global leader in AI. The second one establishes the major objectives for the orientation of researches’ public funding for what is considered to be a “transformative technology with an extraordinary potential for the creation of social and economic benefits”. The strategic plan incorporates seven major investments direction, among them being “understand and address the ethical, legal, and societal implications of AI” or “ensure the safety and security of AI systems” which play a major role in developing the military application of AI. About the latter, the US Administration is engaging, through the strategic plan, the American leadership in the field as a national security objective, to be protected and promoted at an international level. Concretely, starting from the recommendation for the government which “should develop a trans-sectorial strategy about the international commitment in artificial intelligence and elaborate a list with artificial intelligence’s application which would ask for international control and monitorization”, the plan of action establishes the direction according to which “the United State must promote an international environment to support the American research and innovation in artificial intelligence and open its markets to the American AI industry, while it must protect the technological advance in AI and prevent the critical American AI technologies acquisition by strategic competitors and rival nations”.

Also, across the necessary measures for the preparation of AI future, there is the artificial intelligence application in cyber-security and armament system fields. AI technologies should prevent the cyber-attacks by generating threats’ dynamic models in cases of huge, unstable and incomplete data numbers. Furthermore, AI should identify the existence of complex vulnerabilities and implement measures to prevent and counterattack extremely sophisticated attacks. Equally, artificial intelligence should benefit from specific cyber-security measures as well, hence, AI applications and cyber-security should go hand in hand.

Using artificial intelligence in military applications is creating the biggest legal and ethical challenges for the use of AI technologies in the forthcoming wars. Increasing the decisional autonomy of command and control systems and increasing the distance between militaries and danger by filling it with unmanned technique and intelligence ammunition with high-precision, ensured by the use of AI, it is increasing the concerns related to the possibility to maintain the human control over operations’ process and rights’ protection of all human beings inside the fight space, to be directly or indirectly affected by AI technologies’ consequences.

Based on these conceptual fundaments, the executive order of president Donald Trump to maintain the American leadership in artificial intelligence confirms the determination of political and military elites from Washington in preserving the US central role in AI and affirms that “maintaining American leadership in AI requires a concerted effort to promote advancements in technology and innovation, while protecting American technology, economic and national security, civil liberties, privacy, and American values and enhancing international and industry collaboration with foreign partners and allies.” The order established six transversal strategic objectives, available for all the governmental agencies, for ”protecting our technological advantage in AI and protecting our critical AI technologies”. Among these, special attention goes to education and training. AI becomes a priority for the education programmes of all levels, starting with the high school to master programmes and different professional instruction formats. They will offer grants for university research and scholarships for the students whose educational interest is focused on AI. The training programmes in AI of the armed forces will be uppermost financed and it will be encouraged the creation of curriculums for the formal and informal AI programmes.

Protecting the US interests in AI will be coordinated by the president’s work programmes on national security issues and will include classified components for field’s sensitive areas.

President Trump’s executive order ensures the US national initiative and action framework in AI establishes the strategic directions to be followed for the main committees and agencies responsible in the field and eases the attributions of the main decisional points for the development and protection of American AI architecture. Against the more and more complex threats of the current world, AI security dimension has a central role across this architecture, and the responsible institutions for ensuring the national security have a new mission, which is defending United States’ leadership in AI field.  

Defence Department’s strategy on artificial intelligence, between optimism and pessimism against future’s conflicts

Defence’s Department AI strategy, launched on 12th of February 2019, called “Harnessing AI to advance our security and prosperity”, is assuming the intensification of AI adoption to create “a force fit for our times” and starts from defining artificial intelligence as “ability of machines to perform tasks that normally require human intelligence”.

According to this strategy, the American military application of artificial intelligence is justified by the tendency of “other nations, particularly China and Russia, are making significant investments in AI for military purposes, including in applications that raise questions regarding international norms and human rights. These investments threaten to erode our technological and operational advantages and destabilize the free and open international order”.

The strategy offers examples like fields wherein artificial intelligence can have an immediate strategic and operational contribution. Among these, “the improvement of situational awareness and decisional process”, which refers to applications for useful information extraction from primary data and commanders’ equip with situational warning systems to ease the selection of the best operational action method or “improving the safety of operative equipment”, with application in air, naval and land systems safety by alerting operators about the hidden dangers in complex and volatile operational circumstances.

Pentagon’s document does not refer, concretely, to complex high-powered armament systems which could be endowed with artificial intelligence or the way American militaries are speculating the use of artificial intelligence for deterrence and defence against similar systems used by hostile forces. The ineffectiveness of such dates left the door opened for speculations about the replacement of heavy equipment like planes, tanks and ships with softs to have the capacity to make the difference between victory and defeat in futures’ super-technologized conflicts. In that world, complex algorithms will be useful for a large volume of data collected from the operational area with drones and other unmanned systems in order to provide recommendations to commanders in the battlefield or will actually initiate, autonomously, actions dedicated to enemy’s complex armament systems combat, without using lethal force. 

Beyond speculations, it seems that the Defence Department is not that close to creating “killer robots” as suggested in the analysis made by the famous publication Foreign Policy on this matter. Across this analysis, it is quoted gl.lt. Jack Shanahan, who clarifies that the United States “is nowhere close to the full autonomy question that most people seem to leap to a conclusion on when they think about DoD and AI”. For now, the human decision preservation ethic in using lethal weapons seems intangible, even if exists, for quite a while, technologies like “fire and forget” wherein the contribution of the human operator is limited to uploading data about the target before launching, and the missile decides automatically its trajectory, around the targeted objective, based on data collected through radars.

Currently, the DOD Joint Artificial Intelligence Center of the Department of Defence has launched two pilot projects with a moderated ambition level. One refers to assistance in case of disasters, specifically on firefighting, and the second one on predictive maintenance – a new concept for intelligence insurance with exchangeable pieces and intervention for reparations and maintenance of military equipment- for H-60 helicopters. The first project will be made in collaboration with the Homeland Security Department and other governmental agencies, then to be tested the cooperation procedures across the US’s AI national security architecture.

Beyond the concrete projects to affirm the opportunity of AI initiatives funding in the military field, the AI strategy of Defence Department ensures, for the first time, the conceptual framework of complex armament systems development of the future US army. Furthermore, maybe the most important aspect, the document offers continuity in the strategic thinking from Washington, artificial intelligence becoming one of the few fields whose vision and opportunity is not questioned by the current administration, which is still seeking solutions for the effectiveness of national security costs in an extremely unstable period for the international systems since the end of World War II.

Conclusions and AI security dilemmas

For the public, it is still unclear what artificial intelligence means and what exactly can and cannot do artificial intelligence. Employees in different industries are somehow circumspect regarding artificial intelligence. It could ease their job by taking some of their responsibilities and improve their work condition, but it could also replace them for good, leaving them on the streets. Lots of skeptical or radical organizations think that artificial intelligence will not create but social disorder or labor force crisis, but it will even take the globe’s control and will push humanity to the disappearance. However, we should not forget that each industrial revolution in the past two centuries and a half was received exactly the same way, and sometimes it led to fundamental changes, some violent, for the international order, however, the human society was not affected. Why would things be different this time?

The huge potential of AI applications for security and defence is creating the premises of existence in a significant classified space of strategic thinking and AI military projects. There was already leaked information from the inside, which are more or less trustable, regarding unexampled technologies developed in secret laboratories, identified through acronyms and numbers, capable to control life through devices that we use daily not only to communicate, but also for a greater effort while we are at home, job or somewhere else. Therewith, complex algorithms which foresee, prepare and even assume conspicuous decisions over the present and the future would already work with data we use in the virtual space of our existence more or less consciously. Across such a scenario, each human movement or each action or decision of any institution are scanned, tracked and interpreted as the pieces of a global data and information puzzle, whose solutions are satisfying more and more hidden interests.

Regardless of the feasibility of these presumptions, a certainty is that many states have started to organize national systems specific to artificial intelligence against the fourth industrial revolution. Their quasi-wholeness starts from the idea that artificial intelligence will bring huge benefits for the states, hence it must be encouraged and protected even under the aegis of national security priorities and values. However, benefits involve equal costs, either political, economic or social. And when the measure of international success is the state, this division of an eminently all-embracing technology will not lead but to the emergence of  competition and, as always, it won’t be long until it will be transformed into a conflict.

We will see how these costs will influence the architecture of an international system which is already jammed by a lot of perspective and decision breaches on all traditional cooperation and partnership directions, regardless of the position of the Cartesian benchmarks of a policy full or distrust lack of predictability.

Will artificial intelligence, trough its unique anticipatory capacity, be able to overcome the distrust induced to different decisional levels or, on the contrary, through its speculative potential, will intensify the concerns related to allies or enemies’ intentions in crisis situations? The answer can emerge sooner than we have thought. Are were ready for this?

Yearly Review: 2018 Security Agenda in a nutshell

What will 2018 be remembered for, at different layers of security

  • National
  • European
  • Internațional