Artificial Intelligence in the UK: Risks and Rewards
Alexander Karran
2016-05-12 00:00:00
URL



The author thanks members of Transpolitica and the Transhumanist Party UK for their feedback on previous drafts of this report.

Note: click here to access the entire set of submissions accepted by the Science and Technology Committee.



Executive Summary



This briefing introduces Artificial Intelligence (A.I) as it is applied in industry today, and outlines what the United Kingdom can do to take full advantage of the technology. The briefing covers four areas the Implications of robotics and artificial intelligence for the UK, gaining and maintaining primacy in A.I technologies, the social and economic opportunities afforded by A.I technologies and issues in developing robotic and artificial intelligence technologies.





1.    Introduction




  1. Not since the dawn of the industrial revolution in 1750 has there been a period in history that has so radically altered society, economic growth and technological development. Since the start of this revolution, there has been a rapid but linear pattern of growth and development resulting in three distinct eras. The first era was the industrial revolution (mid-18th century), the second was the period of mass industrialisation (mid-19th century), which has now slowed and the third is the Information Technology (I.T) revolution which began in the latter half of the 20th The I.T revolution however, has broken this trend of linear progress and set in motion a period of exponential growth and development. This new revolution, which comes fast on the heels of the previous one, has been termed the fourth industrial revolution[1] or the second machine age[2].

  2. Thus far, this new “digital age” has been characterised by wide adoption of the internet and the creation of so called “cyber-physical” systems -replacing traditional infrastructure with digital technologies- and by convergence, in which the reliance and co-dependence on data driven processes provided by information technologies, is blurring the traditional divisions between scientific and industrial domains. This co-dependence is accelerating scientific and industrial progress in many areas, such as genetic engineering, regenerative medicine and automation driven by advances in robotics and algorithmic control or Artificial Intelligence (A.I).

  3. Advances in A.I have arguably been the principal technology contributing to current progress and one that is evolving in near real-time. A.I is set to become the most advanced technological tool developed by man, since the discovery and use of fire. There are many definitions of artificial intelligence, such as Machine Consciousness, Narrow Artificial Intelligence, Machine Learning, Strong Artificial Intelligence and Artificial General Intelligence. However, for the purposes of this briefing, the focus is upon narrow artificial intelligence applications as these are currently heading into mainstream deployment.

  4. A.I loosely defined, is a set of statistical tools and algorithms that combine to form, in part, intelligent software that specializes in a single area or task. This type of software is an evolving assemblage of technologies that enable computers to simulate elements of Human behaviour such as learning, reasoning and classification. Some examples of A.I include classification algorithms, used to classify images on social media platforms, text mining algorithms and junk email identification to more complex examples used in computational biology and drug discovery, social network analysis, human gameplay and healthcare analytics (such as Google Deepmind and IBM Watson).

  5. In this briefing, we will outline the connotative potential of A.I as a tool to effect radical social change, financial stability and growth and an enhancement or declension of human existence. Our aim is to rationally consider the negative implications of A.I based on available evidence and opinion, while at the same time emphasising the benefits of the technology when ethically applied so that every UK citizen can be given the opportunity to live better than well, which is our mandate.





2.    Implications of robotics and artificial intelligence on the future UK




  1. If the 1st industrial revolution was characterised as a race between technology, labour and education, the 4th may well be categorised as a race against technology that replaces both brains and brawn. This next race has serious implications for education, especially given the current plans to replace state schools with a broad, enforceable national curriculum, with academies that focus on education as product. However, the greatest area of impact will be upon employment, recent research completed by the Bank of England[3] amongst others[4] shows that up to a third (potentially 15 million) of jobs are amenable to automation, by either robotics or software and cover an ever-broadening range of occupations.

  2. The eventual effect of automation will be the creation of an “autonomous economy[5]” in which digital processes talk to other digital processes and synergise to create new processes; this will allow industries to employ fewer workers, yet complete more work. Traditional trend mappings of the economic landscape (i.e. Neo-Capitalism) point to a trend in which, as the automation of labour increase so too does job creation (after an initial period of rapid falloff and recovery). However, there is now a growing body of evidence to suggest that this time things will be both quantitatively and qualitatively different[6], and that this difference is due to wider applications of Moores Law, beyond electronics to machine learning and information science. This new trend presents as an exponential growth curve that will see automation applied beyond physical labour to more cognitive labour[7].

  3. As this trend manifests over the coming decades it will lead to a radical redistribution of the capital-labour ratio by adding a new vector, that of the robot, forever changing the distribution of resources amongst the various strata of society. In certain occupational domains, human labour is likely to continue for technical, economic and ethical reasons. On a technical level, machines today remain inferior to humans at jobs involving creative, highly flexible or affective work and those tasks that rely on tacit rather than explicit knowledge. It may be that it is simply not economically feasible to replace workers in these areas, or ethical as in the case of those providing secondary, tertiary health or palliative care.

  4. The coming changes will lead to a hollowing out of low, mid and high skill workers, who would (based upon previous models) re-deploy to the work force after a period of re-education and skill improvement. However, in the new digital ecosystem, re-education will only take an individual so far, as their ability to adapt to rapid advancements decreases with the increased pace of change. This will require a radical rethink of what it means to learn and work in a rapidly evolving digital economy.

  5. Providing education and income to citizens who have or are likely to lose employment due to automation will prove to be one of the greatest challenges in the decades ahead. To address this challenge and help manage the social and economic impacts, government should engage with the public in open conversation and wide media outreach about the likely impact of A.I. Answering key questions such as how will the nature of work change? What types of jobs are likely to be automated? If there is no work, how will I support myself or my family? What can I do to find work? What will there be for my children? Such answers as provided need to be unambiguous and clear-cut and guide the public towards those industries with less probability of automation and towards education and skills training.

  6. In addition to the implications for the economy, employment, and education, serious consideration needs to be given to the military applications of this technology, in both a global and domestic context. These considerations need to take into account defensive, offensive, automated and autonomous perspectives. International debate is already fuelling calls for a prohibition on the deployment of autonomous systems in the theatre of war on ethical grounds (Red Cross[8], Human Rights Watch[9]). The HRW stated a position with some merit:




“A requirement to maintain human control over the use of weapons would eliminate many of the problems associated with fully autonomous weapons. Such a requirement would protect the dignity of human life, facilitate compliance with international humanitarian and human rights law, and promote accountability for unlawful acts. 






3.    Gaining and maintaining primacy in A.I technologies




  1. In order to gain and maintain leadership in the application and development of A.I technologies the UK must concentrate its efforts to varying degrees in three areas, education, research and industry. The area requiring the highest degree of effort, but potential maximum return will be education. There needs to be a radical reform of the national curriculum that focuses above all else on the creative use of technology from an early age, potentially as early as year two and becoming more intensive by year twelve. In addition to performing “traditional” learning tasks such as reading and writing, technology should be integrated as an additional support tool, allowing other cognitive tasks such as mathematics and the sciences to be reconceptualised to support the step changes in learner outcomes that are required for modern life and the digital workplace.

  2. Traditional educational practice has thus far focused on developing core cognitive competencies, such as reading, writing and arithmetic with very little variation in how these are taught and applied beyond the education environment. Current and in-development A.I technologies shows that machines are already making significant strides towards mastery in all of these areas and it our conclusion that creativity should be added to this set of core competencies, a conclusion already supported by others[12]. Arguably creativity, in the digital workplace, can be defined as the ability to ask questions using advanced intelligent technology and utilise the answers to create novel solutions to present and future problems, as such, a superior proficiency in data analysis to produce insights and apply solutions will be a much sought after skill-set in the coming decades.

  3. Technology and A.I can serve a dual purpose in a reformed educational system by being both the facilitator of high quality learner experiences and the subject of those experiences. If in the UK, we could teach the creative use, support and understanding of A.I and information technologies from an early age, we could potentially create a generation of thought leaders and expert practitioners, with the foresight to fully utilise the potential of A.I both at home and abroad in many industrial domains. However, to achieve this, the educational system would need to foster an environment that encourages critical thinking, rational decision making and multidisciplinary approaches to increase creativity and a learner’s ability to synergise knowledge from disparate sources.

  4. The integration of A.I into education is not something that needs to be developed from first principles, but rather something that needs only draw from the existing field of A.I.Ed[13], a field of research already rich with methods and technologies which given opportunity, funding and study for feasibility could be rapidly deployed. This integration could feed into the current government’s plans for academy style schools, with a call for “match fund” proposals that blend A.I.Ed into current teaching practice and change the classroom environment from one that has barely changed in a century into something representative of modern digital life. This has the potential to rapidly evolve the education ecosystem, as incorporating machine learning into teaching and learning styles, would present an innovative “learning” environment that supports the learner, teacher and the A.I system as it learns how best to serve each learner individually according to their needs, providing a level of personalisation heretofore unknown in teaching practice.

  5. While high quality education concerning A.I technologies is needed at all levels of society, if A.I is to be embraced openly and incorporated fully as part of UK infrastructure, a positive feedback cycle needs to be created between citizens, education, industry and government. If initiated effectively this feedback cycle will fuel growth over and above standard measures of GDP, to include: education as product; innovation and entrepreneurship as a commodity to be shared strategically with allies; digital information infrastructure as a service; and if information security policy interactions are non-repressive, cybersecurity services. Furthermore, if A.I technology development is regulated using a light but firm touch, such a feedback loop allows for both secular development and global participation, providing opportunity for the UK to take global leadership in the development, application and commercial exploitation of A.I technologies.





4.    Social and economic opportunities




  1. The number of social and economic opportunities afforded by developing A.I technologies is practically limitless. The development and broad acceptance of the technology within all levels society will lead to advancements in many disparate fields, from healthcare and healthcare provision; critical infrastructure management and resilience; to decision support tools and forensic process automation. For example, advancements in the field of medicine and biology due to the application of current A.I have been truly revolutionary. A.I has allowed the tremendous amounts of data used in research to be analysed faster than ever before. Future developments in A.I, will further increase the pace of medical discoveries, and lifesaving medical interventions, accelerating discoveries in DNA mapping, drug discovery, genetic modification and synthetic biology, propelling the biological sciences to a whole new level.

  2. The UK as world leaders in synthetic biology and the biological sciences, is well placed to take advantage of A.I technologies in this domain, not just through our research frameworks, but also in the future through a reformed education system that incorporates A.I.Ed and penetrates all levels of society, making the use of A.I to complete tasks or reach goals as natural as using a calculator or pen and paper.

  3. Many pundits, experts, economists and capitalists argue that specialized narrow artificial intelligence applications, robotics and other forms of technological automation will ultimately result in a significant increase in human unemployment and underemployment within many fields of human endeavour (Deloitte[14], Financial Times[15], RSA[16]). This significant “hollowing out” of Labour at all levels of the employment ladder may well result in a fundamental shift in UK society, leading to much greater levels of inequality, lesser social justice and a greater potential for social unrest. However, if managed effectively and with all due ethical consideration, the further development of A.I and the concomitant increase in the automation of labour could become a boon to the UK, in terms of increased productivity by reducing the everyday burden of citizens via a basic-income and freeing their creativity and innate empathy. Additionally, providing a basic income allows for the scrapping of large portions of the welfare system as means testing or fraud detection would no longer be required. This would have the effect of increasing social mobility for citizens at a time when it is required, in order to re-educate or re-train to remain a viable prospect in a shrinking employment market.

  4. Thus, while it is increasingly likely that grande-masse automation will reduce the number of employment opportunities, benefits can accrue through an evolved education system that creates new employment opportunities. A.I.Ed could potentially become a large area of economic growth, helping to increase A.I, robotics and automation acceptance in the general populace. Growth would be stimulated based on the sheer number of industrial and cognitive domains required to support the development of educational A.I systems, such as, educators, computer science, Information technology, designers, technologists, infrastructure specialists, content creators and those sub domains that that support them.

  5. There will be a number of other opportunities both economic and social that will come from developing A.I technologies in the UK and advancements from within the industry. For example, in transport management, imagine a transport management system able to respond in real-time to traffic conditions nationally and locally, with the ability to update automated and non-automated vehicles with prevailing conditions and alternative courses of action, ensuring optimal traffic flow and reducing fuel consumption, accidental damage and time-to-travel as a consequence. The same types of system could be applied to critical infrastructure to provide greater resilience and fault tolerance. Another example within healthcare, would be the use of Big Data analytics to provide diagnostic support to GP’s, hospitals and secondary care providers allowing for fully personalised health care, through the rapid diagnosis and identification of disease states and which interventions would work most effectively for the individual, helping to reduce the cost overhead associated with prolonged care due to diagnostic exploration and drug provision.

  6. The possibility exists that A.I could be used within governance, providing evidence and helping to fact check statements and build policy. A decision support A.I could in effect act as a buffer between politician and policy, ensuring that before policy becomes actionable no unintended consequences are likely to arise. This is also an area of active research that could also provide economic and status benefits should the UK encourage its growth.



5.    Issues in developing robotic and artificial intelligence technologies




  1. The biggest challenge the UK faces with regards to developing A.I technologies, is educating the populace in its use, benefits and risks and how fast this information can be disseminated, from those in early learning, attending college or university to those performing jobs soon to be automated or retired, all must be made aware of the coming changes. It will not simply be a case of injecting some A.I subject matter into schools and colleges and hoping that learners and schools adapt, the change to an “A.I mind-set” needs to be systemic affecting all levels of our society. For the UK to prosper an equal focus must be on the practical applications of A.I in addition to creating and understanding the technology. The UK must focus on creating a generation of machine learning practitioners, through early learning and advocating “degree apprenticeships” or vocational certification.

  2. Short-sightedness could make the UK fall at the first hurdle in its efforts to capitalise upon A.I technologies, existing austerity measures could inhibit any effort through lack of funding. Taking advantage of A.I and it development would require a significant redeployment of funds towards those scientific and industrial domains which demonstrate multidisciplinary approaches utilising A.I to provide services nationally and globally or those applying A.I solve problems specific to the UK and its society.

  3. Another challenge facing the UK will be ensuring positive applications of A.I, a balance must be struck between national security needs and personal freedoms afforded to UK citizens, applying fully autonomous A.I to surveillance tasks targeted at citizens is a minefield of unparalleled danger. While the state is tasked with the security of its people, policing thought and action beyond the confines of just law, lies outside of its remit.

  4. A nimble and lean directorate consisting of ministers, economists, scientists and policy experts and futurists should be created, able to respond in short timescales to technological advances in near-real-time this expert policy group should advise upon and revise policy in line with the pace of technological change. Rather than traditional precautionary policy decision approaches this group should adopt a proactionary approach to policy and regulation (i.e. a light touch, but ethically constrained) of A.I in order to reap the benefits that the technology can bring to society and advance understanding of the negative consequences. This group should advise upon or create policy and legislature that is robust enough to adapt to rapid and radical changes, without falling into traditional deny all regulation.

  5. While it is lamentable that we live in a world of warring nation states, unmitigated threats and intractable ideologies, defence is another area in which the nations technological expertise and thought leadership can be applied. Investment in A.I and robotics for national defence is increasing globally[17], and it is within the UK’s best interest to increase research and development in this area in order to keep abreast of the changing nature of warfare. A full analysis of how A.I can be applied to defence ethically and morally is beyond the scope of this briefing. However, artificial intelligence and automated defence could potentially be an area of economic growth and a driver of global stability for this century, much as nuclear weapons and the potential for Mutually Assured Destruction was for the early portion of the previous century, the potential risks of A.I and robotics applied to warfare cannot be overstated.



6.    Conclusion




  1. A.I research and development has an immense amount of momentum behind it, socially, technically and economically. It is not a question of if we should we develop A.I further but rather how fast can the nation mobilise resources in the industrial, educational and civil services to take advantage of this brief period of research and exploration of the technology. The government must make a statement that defines the nation’s role as a leading light and technologically advanced society to be made to be at the forefront of its development in terms of the nation’s ability to prosper and defend itself.



References

[1] https://www.weforum.org/agenda/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond

[2] Brynjolfsson, Erik, and Andrew McAfee. The second machine age: work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company, 2014.

[3] http://www.bankofengland.co.uk/publications/Pages/speeches/2015/864.aspx

[4] http://www2.deloitte.com/content/dam/Deloitte/uk/Documents/finance/deloitte-uk-finance-robots-are-coming.pdf

[5] https://www.technologyreview.com/s/515926/how-technology-is-destroying-jobs/

[6] Brynjolfsson, E., & McAfee, A. (2014). The second machine age: work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company.

[7] http://www.mckinsey.com/business-functions/business-technology/our-insights/disruptive-technologies

[8] https://www.icrc.org/en/document/statement-icrc-lethal-autonomous-weapons-systems

[9] https://www.hrw.org/news/2016/04/11/killer-robots-and-concept-meaningful-human-control

[10] http://spectrum.ieee.org/telecom/security/the-real-story-of-stuxnet

[11] http://www.bbc.co.uk/news/technology-36158606

[12] https://www.nesta.org.uk/sites/default/files/creativity_vs._robots_wv.pdf

[13] https://www.pearson.com/innovation/smarter-digital-tools/intelligence-unleashed.html

[14] http://www2.deloitte.com/uk/en/pages/press-releases/articles/automation-and-industries-analysis.html

[15] http://www.ft.com/cms/s/2/126527ce-f8a8-11e5-8f41-df5bda8beb40.html#axzz45QbZXYUO

[16] https://www.thersa.org/discover/publications-and-articles/reports/basic-income/

[17] http://www.ft.com/cms/s/0/9067b4c2-0d14-11e6-ad80-67655613c2d6.html