Written evidence submitted by Innovate UK (ROB0060)

1. Innovate UK is the UK’s innovation agency, a non-departmental public body sponsored by BIS.  It is the prime channel through which the Government incentivises innovation in business.  Innovate UK is business-led.  Our governing board and executive team is comprised of experienced business innovators and experts.  We work with people, companies and partner organisations to find and drive the science and technology innovations that will increase productivity and exports and grow the UK economy.

2. We are working to:

• accelerate UK economic growth by nurturing small high-growth potential firms in key market sectors, helping them to become high-growth mid-sized companies with strong productivity and export success;
• build on innovation excellence throughout the UK, investing locally in areas of strength;
• develop Catapult centres within a national innovation network, to provide access to cutting edge technologies, encourage inward investment and enable technical advances in existing businesses;
• turn scientific excellence into economic impact and deliver results through innovation, in collaboration with the Research Community and Government; and,
• evolve our funding models to explore ways to help public funding go further and work harder, while continuing to deliver impact from innovation.

3. In line with our strategy[1] we operate across Government and advise on polices which relate to technology, innovation and knowledge transfer.  We also support Government departments to become more efficient by supporting them in developing innovative solutions through harnessing the creativity that businesses can offer.

4. Innovate UK was established in July 2007 (as the Technology Strategy Board). We have invested over £1.8 billion in innovation, and have helped more than 7,600 innovative companies in projects estimated to add up to £13.1 billion to the UK economy and created an average of 7 jobs per company we have worked with. Our investment over the last 8 years has meant that every £1 invested has returned up to £7.3 GVA to the economy and created 55,000 jobs. The private sector more than matches that investment, doubling the power of public sector money. We work with nearly every University in the UK to stimulate the commercialisation of leading-edge academic research and innovation.

5. Driving productivity and growth is at the heart of Innovate UK’s strategy and purpose, so carefully considering the future economy is key to setting our direction.  Innovate UK welcomes the fact the Science and Technology Committee is undertaking an inquiry into robotics and artificial intelligence. Our Emerging and Enabling Technologies programme seeks to identify, and invest in, technologies and capabilities that will lead to the new products, processes and services of tomorrow – those with the potential to create billion-pound industries and disrupt existing markets. Our Enabling Technologies programme will concentrate investment on innovative application and service-led solutions in the areas of: cyber security; data; satellite earth observation; electronics, sensors and photonics; the internet of things and; robotics and autonomous systems. Set out below is our response to the questions raised by the Committee.

1. The implications of robotics and artificial intelligence on the future UK workforce and job market, and the Government’s preparation for the shift in the UK skills base and training that this may require.

6. We’d like to begin by defining some terms so that our submission may be seen in the intended context.

Robotics

•                               Robotics is the branch of science and engineering & technology that deals with the design, construction, operation, and application of robots/robotic systems, which in turn are machines capable of carrying out a series of actions on behalf of humans.
•                   Until recently, robotic systems were mostly used in industrial applications and academic research. These systems are enabled by advances in embedded computation, communication, and robot-associated software. They enable the real-time interpretation of the physical world, combined with real-time controlled response and reaction, with or without a human operator being involved. These are traditional programmable systems based on rules that process data through a series of predetermined steps/procedures to arrive at outcomes based on structured data.
•                   Robotics has increasing overlaps with electronics, computer science, mechatronics, nanotechnology, bioengineering and other disciplines.
•                   Currently industrial robotics is the dominant market sector. Such robots perform physically difficult, dangerous, or repetitive tasks that would otherwise have to be carried out by by humans, and they tend to be expensive, mostly bulky, and execute specific and, usually, a limited range of tasks.
•                   Today’s more advanced robotics systems strive to incorporate additional functionality (such as re-configurability, adaptivity, and dependability/reliability), and have enhanced abilities (sensing, manipulation, mobility, interactivity) and, in some cases, seek to resemble humans in appearance and behaviour (humanoids).
•                   Generally, the powerful and complex software that these robotic systems run remains “static” (i.e. their logic does not change) during their operation.

Artificial Intelligence (AI)

•                   There are many different definitions of artificial intelligence (AI), but we consider AI to be the bringing together of science, engineering and technology with the objective of making a computing system, or machine, mimic human behaviour in many broad and varying contexts.
•                   Many different technologies need to be brought together to make an AI system function, including, depending upon the use case: sensing (including navigation, computer vision, 3D sensing, situational awareness etc.); natural language processing; reasoning; machine learning; knowledge representation; planning; and higher level cognition or intelligence (strong AI)
•                   AI technologies/systems are often also called cognitive technologies/systems (e.g. cognitive computing).
•                   A large part of current AI work focuses upon machine learning, which entails the design and development of computing systems and applications capable of learning based on their data inputs/states/outputs, i.e. learning by experience. Examples where machine learning systems perform better than humans in specific tasks already exist; Google’s AlphaGo computer that recently beat the Korean grandmaster Lee Sedol was based to an extent on machine learning technologies.
•                   Whilst true “artificial intelligence” technology is still considered by many as futuristic, the closely related fields of data fusion, analytics, mining and machine learning technologies are already finding their way in many different applications and demonstrating high value.
•                   Currently, AI implementations operate in very specific contexts (playing games such as jeopardy or chess, providing certain diagnoses based on medical images, learning how to grab boxes from a pallet, etc.).
•                   The main difference between the AI/cognitive systems and the “traditional” programmable ones is that, in the former the logic process it is not static and can become modified during operation, based, for example, on data and experience; they are designed to adapt and make sense of the complex and unpredictable unstructured information, often assigning confidence levels to potential insights or answers.

Autonomous systems

•                   Different definitions exist for autonomous systems; in most cases the view taken for them is one of the external observer, and refer to systems that operate without a Human-in-the-Loop during the execution of their tasks.
•                   While this definition can serve many purposes, it does not distinguish between systems that have “operational” autonomy, via some type of closed loop control with already embedded control laws or logic (which can also be called automated or adaptive systems[2]), and systems which can “learn” and thus create their own logic, situation awareness, planning capabilities and their “own” laws or logic, thereby enabling “decisional” autonomy.[3]
•                   While the terms automated and autonomous tend to be used interchangeably, in order to minimize confusion we adopt the automated vs autonomous distinction herein; the automated systems do not include AI technologies like reasoning, learning, knowledge representation, planning or other higher level cognition, only the autonomous systems do.
•                   Automation and autonomy can be thought of as comprising the two ends of a spectrum.
•                   Current autonomous systems contain a set of cognitive capabilities allowing them to operate within certain situational boundaries.
•                   The vast majority of current robotics systems are automated. Driverless/self-driven vehicles, that are often called autonomous vehicles, are also referred to by the corresponding industrial or regulatory bodies as ‘automated vehicles’.
•                   Autonomous and/or automated systems may operate in the digital/cyber domain or, with the addition of sensors or actuators, within the real, physical world (e.g. autonomous robotics). 

7. The implications on the future UK workforce and job market

Both robotics and artificial intelligence technologies will have an impact on the future global workforce and job market.

8. Robotics: the International Federation of Robots (IFR) estimated that the total world market for robot systems in 2012 was $26 billion, and, based on a study by Metra Martech "Positive Impact of Industrial Robots on Employment”, the one million industrial robots currently in operation have been directly responsible for the creation of close to three million jobs. A report by McKinsey[4] estimated that by 2025 the application of advanced robotics could generate a world-wide  economic impact of $1.7 trillion to $4.5 trillion per year, and the use of advanced robots for industrial and service tasks could take on work that could be equivalent to the output of 40 million to 75 million full-time equivalents.

9. Industrial robotic systems have been designed, developed, manufactured and used in the UK for decades. However as the technology of advanced robotics develops, a more highly skilled workforce will be needed to undertake the design, programming, assembly and maintenance of these more complex systems.

10. The UK is already one of the world’s leading hubs for robotics research, with a long history in the field. The UK is also home to many world-class companies, in different sector of the economy, including many supported by Innovate UK. We now have the chance for these companies to benefit from the opportunities next generation robotics and autonomous systems (RAS) present, to increase their productivity and create new products and services.  A 2014 survey by EEF, the Manufacturers' Organisation, indicated the adoption of robotics, with the corresponding improvements in productivity in manufacturing would also lead to re-shoring to the UK.

11. There is the potential for the UK to take a lead in the next generation of advanced robotics, particularly in the service segment, the global market for which is expected to grow from 7 Billion USD in 2014 to 18 Billion USD by 2020, representing a CAGR of 17.4% between 2015 and 2020[5]. A sensible aspiration would be for the UK to capture perhaps 10% of this market.

12. The net effect on the total number of UK jobs is hard to foresee, but the trend is likely to be towards increasing productivity and the value associated with each job.

13. Artificial intelligence technologies: the same 2013 McKinsey report estimates that the economic impact of automation of knowledge work (through the combined advances in computing technology, machine learning, and natural user interfaces) by 2025 will be in the range $5.2 trillion to $6.7 trillion per year world-wide.  According to a recent report by Tractica[6], which examines the practical application of AI within commercial enterprises, the global market for enterprise AI systems (excluding the additional associated investments or expenditures in professional services, and ICT hardware and services) will increase from $202.5 million in 2015 to $11.1 billion by 2024. 

14. Different scenarios are analysed in “The Future of Work - Jobs and Skills in 2030,” study by the UK Commission for Employment and Skills in 2014, and provide a good framework for thinking about future scenarios for jobs, skills and sectors in the UK; the following four scenarios of the future of work are elaborated:

• Forced Flexibility (business-as-usual) Scenario: greater business flexibility and incremental innovation lead to modest growth in the economy, but this flexibility often results in fewer opportunities and weakened job security for the low skilled.
• The Great Divide Scenario: despite robust growth driven by strong high-tech industries, a two-tiered, divided society has emerged, reinforcing the economic position of the “haves” and “have nots”.
• Skills Activism Scenario: technological innovation drives the automation of white-collar work and brings large-scale job losses and political pressure, leading to an extensive government-led skills programme.
• Innovation Adaptation Scenario: in a stagnant economy, improved productivity is achieved through a rigorous implementation of ICT solutions.

15. Artificial intelligence and robots are identified as one of the ten disruptions that could radically change the future of work in UK[7]. Different views exist on the impact of advances in robotics on jobs, with both scientists and economists offering wildly varying views for how deeply automation will affect future employment.  For example, in “The Rise of the Robots: Technology and the Threat of a Jobless Future“ by Martin Ford, the main scenario is one where advances in robotics could wipe out jobs and deepen inequality. According to Andy Haldane, the Bank of England’s Chief Economist, 15m jobs in the UK are "at risk of automation" by smart machines over the next two decades.

16. Although many jobs are likely to be threatened by new RAS technologies, history (for example the agricultural and industrial revolutions) tells us that new technologies also create entirely new forms of employment that simply did not exist before. Although previous technologies have always resulted in a net gain in employment, there is debate about whether this generation of technologies will create the same outcome.

17. That said, two recent studies by Deloitte [8],[9] concluded that the UK is benefiting from the recent technological changes in terms of employment and the continued success will depend on the ability of businesses, educators and government to anticipate future skills requirements and provide the right training and education.

2. The extent to which social and economic opportunities provided by emerging autonomous systems and artificial intelligence technologies are being exploited to deliver benefits to the UK.

18. The UK derives huge benefit from the use of industrial robotics (in our manufacturing base, particularly automotive), although the supply of these traditional robotics systems is mainly from Germany and Japan.

19. In more advanced robotics systems, linked to AI or autonomous systems, although the potential is high, our view is that the extent of successful commercial adoption is currently relatively low worldwide, including in the UK. Much technology remains to be developed and there are other issues to be resolved before widespread adoption can take place.

20. Like all new technologies, RAS will go through a series of stages before it eventually becomes mainstream. Towards this, the following issues will have to be addressed:
•                   Public perception: some science fiction writing has raised concerns in the mind of the public on the safety of such systems. For example, a key barrier to public acceptance of driverless cars is the lack of confidence in how they might respond to the real events that happen on real roads, such as interaction with cyclists, pedestrians and other drivers. These concerns, combined with the fear of potential unemployment, require some thoughtful public engagement to be conducted.
•                   System resiliency, cyber-hardening, cyber-security: concern over the vulnerability to hacking of RAS devices should be taken seriously and systems to prevent unauthorised access or modification put in place.
•                   System verification & validation and appropriate accreditation/certification: no clear paths exist for the verification and validation of autonomous systems whose behaviour changes with time, as they learn and adapt during their deployment.
•                   Legal aspects of responsibility, liabilities etc.: Allocating responsibility, blame and costs/liabilities are issues yet unresolved for autonomous systems.

21. For these reasons, it is expected that for the foreseeable future commercial applications will retain a human “in the loop” to oversee, guide or improve the use of autonomous systems.

The above notwithstanding, the UK is well positioned in the area of remotely piloted autonomous systems (RPAS), including through the financial support provided by BIS / Innovate UK (then Technology Strategy Board) through the ASTRAEA programme, a £62m investment between 2006 and 2013. The UK is also well positioned to benefit from future autonomous vehicles markets, including platooning, although this is still an emerging area.

22. In other future application areas the UK is less well poised (see below), but there have been isolated instances of remarkable success, most recently Swiftkey, who created a machine learning algorithm that produces more accurate predictive text. £65,000 of grant funding to SwiftKey from Innovate UK helped to develop their predictive text App for Android and iOS devices, such as smartphones and tablets. Swiftkey recently completed a $250m exit to Microsoft.

3. The extent to which the funding, research and innovation landscape facilitates the UK maintaining a position at the forefront of these technologies, and what measures the Government should take to assist further in these areas.

Opportunities in RAS

23. It is clear that many of the large companies that will provide a route to market for robotics technologies will need to draw on SME and academic talent to integrate RAS technologies into their product and service offerings. Similarly, SMEs and start-ups need to be encouraged to develop new supply chains for RAS products and services.

24. The risks of not investing in RAS are real, if we do not invest in RAS we run a risk that we lose the opportunity to develop a vibrant and successful RAS industry in the UK and we will lose out to international competitors. Our present UK advantage resides in our creativity, academic strength and engineering capabilities but the window is closing. By investing in the RAS innovation pipeline, the UK has the potential to bring wider operational and commercial benefits to industry, such as improvements in safety (particularly reducing the risks in hazardous environments) and reduction in environmental footprint.

25. The market potential of RAS technologies is significant.

RAS technologies are estimated to have a potential global economic impact of $1.9 - $6.4 trillion by 2025. The UK is primed to take advantage of this opportunity and increase our market share[10]. RAS technology has the potential to increase the productivity of other sectors of the UK economy by up to £218 billion (15% of GVA)[11]. The Electronics Systems Council (ESCO) has prioritised RAS in their work to grow the UK’s electronics industry to £120bn by 2020. Electronic systems are a critical enabler of RAS applications across multiple market sectors and will support key supply chains.

26. Other countries around the world have noted the opportunity, and have been making significant investments on behalf of their national economies. For example, Korea invested $100 million per year for 10 years (2002-2012) in robotics research and education as part of their 21st Century Frontier Program; Japan is investing $350 million over the next 10 years in humanoid robotics, service robotics, and intelligent environments; and in the US, there is a National Robotics Initiative (£24m in FY2013) and other efforts supported by Defence and regional funds as well as private investment, but so far as we are aware no nationally coordinated US programme exists.

27. With the exception of the ASTRAEA programme, and the Centre for Connected and Autonomous Vehicles (CCAV, see below), Innovate UK has never run a dedicated programme in RAS technologies. That said, we have made isolated investments in these technologies where innovators have come forward in response to competitions run in other areas and proposed RAS solutions. Examples of our investments are given below:

• Independent living:  The Office for Budget Responsibility estimates that by 2065, 26% of the population of England and Wales will be more than 65 years old, up from 18% today. Through the CHIRON project, a consortium of leading researchers, care providers and robotics experts have received a major award of over £2m from Innovate UK to develop a modular robotic solution for the home, to revolutionise long-term care by giving people the choice to stay independent in their own homes as they age.

• Driverless cars: Innovate UK has initiated a programme of research, development, demonstration, and deployment of connected and driverless vehicles including: publishing the Department for Transport’s (DfT) code of practice for testing driverless cars; launching the Department for Business, Innovation and Skills’ (BIS) £20 million feasibility studies and collaborative research and development competition Connected and Autonomous Vehicles, the winners of which were announced in February 2016; working with the newly established a new joint policy unit – the Centre for Connected and Autonomous Vehicles (CCAV). CCAV’s remit is to help ensure that the UK remains a world leader in developing and testing connected and autonomous vehicles by leading innovating policy development in this sector; delivering a programme of research, development, demonstration, and deployment activity, worth up to £200 million, through Innovate UK; providing co-ordination across DfT, BIS and the rest of government; and being the single contact point for stakeholder engagement.
• Software Verification & Validation for Complex Systems: Robotics and autonomous systems have complex emergent behaviours, and Innovate UK together with DSTL recently awarded £0.5m of investment to successful proposals to address challenges in this area.
• Artificial intelligence: Even though we have not run a competition specifically on Artificial Intelligence, we have funded proposals to develop the use of artificial intelligence and related technologies for data and user experience in different market applications.

28. Although public sector innovation investment outside of the transport sector has been low in the UK, there has been a healthy investment in the research base. The Government provided £35million of RAS capital investment in the 2012 Autumn Statement to academia, which has leveraged a further £14.5 million, including from industry. It also invested £18.6 million in four new Centres of Doctoral Training in RAS, leveraging a further £20 million from private and Higher education sources.

29. Strategic investments so far have modernised our research institutions with state of the art robotics equipment and strengthened the research and training capabilities in the UK. To harness the critical mass generated through these investments and and fulfil the recommendations of the RAS strategy we need to go further. A coordinated RAS programme will empower RAS SMEs as inventors and innovators; deliver clear returns through multinational end users; generate direct equity investment; build RAS clusters between research and businesses; deliver UK economic growth including social gain from delivery of services; and enable the general public to adopt and use new innovative RAS technologies.

30. Whilst it is very important to support early stage research, if the intention is to grow the UK economy it is equally important to support the emerging industry, including SMEs that form critical parts of the supply chains that will bring products and services to market.

31. Measures the Government should take to assist further in these areas

There are three main things the Government could do, in our opinion, to strengthen investment in the area of robotic and autonomous systems, as timing and funding permits:

• Invest in innovation programmes, supporting companies to develop RAS technologies to a maturity where they are internationally leading and ready for adoption.
• Invest in innovation infrastructure, through a RAS Catapult, to provide the critical mass of underpinning expertise in hardware, software and systems.
• Invest in innovation environment, supporting activities essential to develop the sector, such as standards, regulation, addressing liability issues, public engagement and facilitating access to finance.

32. A coordinated programme of investment from Government allocated to RAS, and matched by Industry, will support the UK’s innovation landscape to drive market growth. It will respond to the immediate needs of industry in priority sectors, and build agility for the future by running targeted RAS grant-funding competitions involving both Industry and Academia. This will ensure the maximum collective value is extracted from the Government’s investments in RAS science and aligned areas, such as Internet of Things and Big Data, building on the excellence of the UK’s research base.

33. A RAS Catapult (working closely with the academic RAS Institute) will bring together in one place a critical mass of the differing types of expertise and equipment (software, hardware, integration capabilities etc.) needed to help UK companies develop commercially viable RAS products and services. It will provide a unique national resource and common point of reference for the regulation and standardisation vital to the stimulation of supply chains. The catapult will provide state of the art test and validation facilities for hardware, software and the cross sectorial capabilities necessary to build effective value chains and linking them to end-users. It will coordinate existing ‘distributed hubs’ where each hub stimulates development as a ‘smart specialisation’ bringing EPSRC and Innovate UK investments closer together. It will also support end-to-end RAS product /service developments to encourage investment in start-ups, collaboration with SME’s, Large Enterprises and Academia.

34. Investments in the innovation environment will provide a coherent national focal point for market-led RAS activity, presenting a visible and open front door to engage end users and international inward investment. It will facilitate effective business-to-business, and business-to-academia collaboration and promote genuine customer engagement with and within supply chains to help identify and overcome commercial challenges. It will communicate effectively across both national and regional government, inform policy and drive the creation of an effective regulatory environment.

4. The social, legal and ethical issues raised by developments in robotics and artificial intelligence technologies, and how they should be addressed.

35. RAS has the potential to provide many positive benefits: to create high value jobs in both the development and use of these new technologies; to improve workplace safety in high risk areas such as inspection or nuclear; to enhanced quality of life by providing solutions to, for instance, the care market; and to enhance national wealth through the creation of new industries or increasing the productivity of existing ones.

36. To realise these benefits many new highly skilled and high value jobs will be created, not only in engineering but also in software, management, creativity, entrepreneurship and complex problem solving. In some instances, this may elevate low-skilled and routine jobs to higher-skilled roles. But there is also a risk that some roles become redundant. Navigating a path through this transition raises a number of obvious social issues.

37. Appropriate legal and regulatory frameworks will have to be developed to support the more widespread deployment of robots and, in particular, autonomous systems[12].  Frameworks need to be created to establish where responsibilities lie, to ensure the safe and effective functioning of autonomous systems, and how to handle disputes in areas where no legal precedence has been set.  This will be essential to help UK companies to deploy some of the technologies they have developed. The UK is mindful of these issues, and we are well positioned in the area of automated vehicles, as recently highlighted in the Department for Transport’s Pathway to Driverless Cars Summary report and action plan[13].

38. The development of robotic systems can also raise ethical issues. This was first highlighted in the context of weapons development for the military. Although there are still no completely autonomous weapons systems, the trend towards more and more autonomy in military systems is clearly visible. This raises issues of international regulation and control and has attracted interest from bodies such as Human Rights Watch[14]. Currently the ultimate authority for launching a weapon from a robotic system lies with a person (who always remains in the loop).

The possibility that AI robots might develop consciousness, and even their own ethical frameworks, remains in our view some way off. However, public concern over RAS is often due to speculation, sometimes inspired by science fiction, on where technology development may end up.

39. When deploying technologies that access personal or private information, there are guidelines (under the UK Information Commissioner’s Office) on how to use and manage the privacy and consent of personal data. General issues for privacy and confidentiality; respect for human dignity and human rights; respect for cultural diversity and pluralism; environmental guidelines and measures; and sustainable development all need to be taken into account when developing such robotics and/or autonomous systems. British Standards Institute identified these issues, amongst others, in standard BS 8611, “robots and robotic devices Guide to the Ethical Design and Application of Robots and Robotic Systems”.

40. Research undertaken by universities, institutes or independent research organisations should follow good scientific practices in accordance with Research Council policies. Research projects that raise potential ethical issues must first seek approval from their institutions’ own committees, and academics applying as part of a business-led consortium have a duty to highlight potential ethical issues via the JeS form. Innovate UK has developed a Responsible Innovation Framework, to help guide work in emerging technology areas as they move from R&D to commercial deployment. Responsible innovation requires anticipation, reflection and deliberation around the purposes of the research, its potential benefits, and the relevant ethical, societal and regulatory issues and appropriate response throughout the process, including (i) during the process of carrying out the R&D, and (ii) for commercial use of the findings. Innovators undertaking work in the area of RAS might find this frame useful.

May 2016