AI in the UK: ready, willing and able? Contents

Appendix 4: Historic Government policy on artificial intelligence in the United Kingdom

The Committee appointed Angelica Agredo Montealegre, a PhD student at King’s College London, as a Specialist Adviser to conduct research into historic Government policy on artificial intelligence. The following is the result of that work.

This note presents a long-term perspective on artificial intelligence (AI) research and development (R&D) in the United Kingdom in the last 40 years. In particular, it outlines the results and implementation problems of Government-supported AI R&D projects.

The note is divided into three main parts. The first part provides an overview and shows that the way AI has been understood in the past 40 years has varied considerably, and that it has usually been considered as an aspect of information technology (IT) policy, rather than an entirely distinctive field. The second part presents the Alvey programme (1983–1987)—the only large-scale national project of the sort during this period—and outlines its objectives, achievements and execution problems. The final part draws a comparison between the Alvey programme and its counterparts in Japan, Europe and the USA.

Overview

The potential of AI generated great enthusiasm and high expectations in the 1950s, leading to the formation of a number of major AI research centres in the UK at the universities of Edinburgh, Sussex, Essex and Cambridge in the 1960s. But by the 1970s this enthusiasm had begun to wane, as promises went unfulfilled, both in the UK and in the USA, the other major global centre of AI, and members of the research community became embroiled in fierce disputes regarding the nature and aims of AI research. This discord, along with the broader backdrop of disappointment, prompted the Science Research Council to commission an inquiry into the state of the field. The resulting report, produced by Professor Sir James Lighthill in 1973, made clear his pessimism about the potential outcomes of basic research in AI in the near future.586 The Lighthill Report is now mostly known within the AI community for causing a reduction in support for AI research in the UK, a period also known as the first ‘AI winter’.

The 1980s saw the creation of the Alvey programme (1983–1987), the first large-scale R&D project involving AI in Britain. Before this programme there were no large national projects featuring AI; instead the Government funded AI through the Science Research Council at universities such as Edinburgh and Cambridge. The launch of the Alvey programme was a response to the creation of the Japanese Fifth Generation Computer programme in 1982. After its success in the electronic consumer goods and automobile industries, Japan announced its attempt to create a new computer with the capacity to solve problems on its own. In order to facilitate its use, the creation of this computer was accompanied by developments in the human-machine interface (in the same way that virtual assistants such as Siri and Alexa are attempting to do today). In this context, the Alvey programme was intended to establish Britain as a key player in the IT sector worldwide.

However, by the late 1980s, the UK’s IT sector had built up a considerable trade deficit, and by the early 1990s it was considered unlikely that the Alvey programme would lead to any substantive commercial returns.587 The evaluation commissioned by the Department of Trade and Industry after the programme stated that the Alvey programme’s focus on pre-competitive research was one of the main factors that hindered its contribution to the enhancement of the UK’s competitiveness in the IT market.588 The Government subsequently rejected proposals to create a follow-up project and the aftermath of the Alvey programme is sometimes referred to as the second ‘AI winter’, which lasted until the early 1990s. This coincided with a general loss of enthusiasm for AI in the US.589

In the UK in the 1990s, support for AI research stopped being systematic, and there were no attempts to co-ordinate research at a national scale. Instead, the UK’s participation in the European research project ESPRIT was intensified.590 Even though no national large-scale programmes were created in the UK, AI research continued, both in industry and in the same universities which had led the way in the early 1970s—albeit at a reduced scale. More recently, emphasis shifted towards privately funded research. As such, the Alvey programme represented the first and last major government-funded AI project in the UK.

It should be noted that a lack of clarity in terms of definitions and objectives seems to have plagued the field right back to its origins in the 1950s. This makes tracing the evolution of the AI field in the UK a difficult task. It is usually unavoidable to refer to IT in general and, even then, information is scant. For instance, the governments of the day appear not to have collected comprehensive, systematic data related to the amount of funds spent on IT R&D, let alone specifically on AI. As such, the best source of information that is available about AI development in late twentieth-century Britain relate to the Alvey programme. Indeed, after the Alvey programme, AI R&D policy was spread between different funding councils and stopped being systematic, making it very difficult to trace.591

Defining artificial intelligence and its aims

The definition of AI has changed over time, and variations on the meaning can be found within the same period—as AI pioneer John McCarthy once put it, “as soon as it works, no one calls it AI any more”. This lack of precision has historically generated confusion amongst policy makers and members of the IT community—as it still does today. Indeed, in the past, there was no consensus about what the goals of AI research should be, or how best to achieve them. This is a crucial issue because the Government, researchers, suppliers and users have had different ideas of what the role of AI is, as well as its potential, contributing to disappointment from different parties when these expectations have not been met.

In the 1970s, there was no consensus amongst researchers about the definition of AI, its objectives, or its economic potential in either the short or the long term. In his report, Professor Sir James Lighthill attempted a description of the AI field that generated various criticisms. Professor Lighthill divided the field into three categories: advanced automation, computer-based central nervous system research, and “robot-building”.592 The objectives of these distinct areas were, respectively:

Professor Lighthill stated that research in AI would probably not yield significant achievements in the following 25 years, and that the creation of an intelligent, general-purpose system was a goal that would not be fulfilled in the twentieth-century—if ever.594 However, he did point out that the chances of success in AI would increase if research was integrated with the field of application.595 Professor Lighthill’s categorisation and conclusions were contested by leading figures of the field in the UK, such as Professor Stuart Sutherland (founder and head of the University of Sussex’s Laboratory of Experimental Biology), and Professor Donald Michie (director of the University of Edinburgh’s Department of Machine Intelligence and Perception). They suggested replacing the category of “robot-building” with basic research.596 They were also more optimistic about the field’s achievements and promises, and stated that a large investment in basic AI was justified if Britain’s AI field was to be competitive worldwide.597

In the 1980s, the UK AI research community coined its own term for AI: Intelligent Knowledge Based Systems (IKBS). IKBS described “systems which combine hardware and software in order to achieve the goal of using inference to apply knowledge to perform a task”.598 However, ‘AI’ was still used, especially when referring to ‘Expert Systems’—computer systems that attempt to emulate the decision-making of a human by using a set of facts and rules that had been previously programmed into them. Even within the category of ‘Expert Systems’ there was no consensus amongst researchers about the aims of the field, which caused delays in the acceptance of research grants (as was the case, for instance, at the Medical Research Council).599

In terms of wider R&D policy, the Government usually considered AI as an aspect of IT policy, rather than an entirely distinctive field. Indeed, it should be noted that AI (or IKBS) was only one aspect of the Alvey programme, which meant that the benefits expected from the project were expected to be advancements in the IT sector, not just AI.

The Alvey programme

The Alvey programme was a five-year collaborative R&D programme in IT which began in 1983. It was funded by the Department of Trade and Industry (DTI), the Science Engineering Research Council (SERC) and the Ministry of Defence (MoD). In total the programme cost £350 million (approximately £1 billion today), of which £200 million came from the Government and the remainder from industry.

Not all funding was new. For instance, the Very Large-Scale Integration (VLSI) sub-programme was based on a similar programme already planned by the MoD.600 Over 300 sub-projects were initiated, nearly 200 of them involving both industrial and academic research teams. The remainder, about 8% of the project budget, were smaller academic-only sub-projects in which companies took an interest. The research strategy was determined by a Directorate, staffed partly by industrial and academic secondees, which oversaw the formulation and implementation of the programme strategy.

Objectives and achievements

The programme had three main categories of objective:

Upon the conclusion of the Alvey programme, participants deemed the technological objectives to have been met. They believed that the programme had correctly identified and supported certain enabling technologies, which remained critical at the beginning of the 1990s. The programme as a whole was primarily focused on producing software rather than hardware.602 The achievements in IKBS were considered to be lower than in other areas of the programme.603

Following the end of the programme, the DTI commissioned an evaluation of its achievements to be carried out by the Science Policy Research Unit at the University of Sussex, and the Programme of Policy Research in Engineering, Science and Technology at the University of Manchester. These evaluators deemed the structural objectives to have been successfully met. The number of researchers increased, with over 5,000 people involved in the programme.604 Moreover, the quality and extent of communications and links between the members of the growing IT community were improved. However, no new research centres of any significance emerged after the programme. The Alvey programme probably benefitted from the existence of important research centres, such as Cambridge, but did not appear to have made significant changes to the size or distribution of research activities in the UK.605

Despite these apparent successes, the strategic objectives were not achieved. By 1980, the UK IT sector had a trade deficit of £300 million, and while the directors of the programme projected it would reach £1 billion by 1990, in reality it was surpassed as early as 1984.606 Although it was expected that in the long-term the programme’s work could lead to commercial returns, by the early 1990s it was considered that these expectations were unlikely to be met.607 Although it was never made explicit what was meant by ‘long-term’, the time ranges given were usually of 10 years or more.

According to the evaluators, the obstacles for the commercial exploitation of the Alvey programme’s work were similar to the ones present in any collaborative R&D project at this time. Firms often changed strategy, which affected their partners’ expected exploitation channels. Moreover, inadequate management of the connections between R&D and production was a common issue, which was further exacerbated by capital shortages. Collectively, these all worked to hamstring attempts to move from research to production.608

Implementation problems of the Alvey programme

During implementation the programme encountered a variety of problems. The director of the programme from 1983 to 1987, Brian Oakley, identified the difficulty of financing the academic part of the programme as the most challenging issue he had to deal with. In his view, academic funds were over-allocated, which resulted in cash flow problems.609

Moreover, Mr Oakley pointed out that many research assistants working on the Alvey programme were not from the UK, and most subsequently returned to their country of origin at the end of their contract.610 He saw this as a problem because skilled staff trained by the programme were not subsequently benefiting the UK. Mr Oakley suggested that several research assistants had probably left academia to join industry because of higher salaries, but he lamented the lack of data collected on the subject. On a more positive note, be observed that some assistants had helped organise IKBS schemes at the Glasgow Turing Institute and at Imperial College, which had provided opportunities to graduate students to work with AI experts for six months.611

A related problem was the lack of qualified staff in both industry and academia. The scarcity of staff was frequently mentioned as a persistent issue throughout the programme and the evaluators in 1991 argued that education and training activities complementary to the Alvey programme should have been put in place to alleviate the situation.612 Neither the directors nor the evaluators of the programme discussed the reason for this scarcity of qualified staff. The evaluators of the programme suggested that the lack of qualified staff was to be expected given the novelty of the research field.613

In many ways, this shortage was surprising, as Britain had led the world in electronic computing during, and after, the Second World War, and it appeared that staff shortages only became prevalent during the 1970s. The historian of technology Dr Marie Hicks has argued that the British government’s persistent computer labour problems were in great part a result of the continued neglect of women’s labour from the 1960s onwards.614 Women had dominated the computing workforce in the initial post-war decades, when many positions were considered essentially clerical in nature, and therefore had less status attached to them. But by the 1970s, as computing jobs shifted from machine operations towards higher level management and strategic roles, women were increasingly passed over, even as vacancies at the top continued to grow.615 The female workforce gradually concentrated in lower-paid and part-time employment and by the mid-1980s, 45% of all women employed in computing were hired only part time.616

Despite Mr Oakley’s claims that the Alvey programme contributed to the expansion of trained staff, by the 1990s the shortage of skilled staff remained a problem.617 If indeed it was the case that a large number of researchers in the Alvey programme were from overseas, this raises further questions as to why the UK was unwilling or unable to retain them after they completed their contracts.

A further source of problems for the Alvey programme was related to the support it attempted to provide for the British computing industry. Although the programme was meant to contribute to the UK’s competitiveness in the sector through pre-competitive research, measures were taken to advance this goal more directly. This assistance took two forms: capital purchases and small firm participation. At the beginning of the programme, the Alvey Directorate decided to make bulk purchases of British computing equipment. Buying the same equipment for the participants was meant to facilitate communication between different research projects.

However, not everyone involved with the Alvey programme agreed with the decision to buy British equipment, with some voicing concerns about quality. Even Mr Oakley later admitted that these purchasing decisions were, in hindsight, “an expensive mistake”.618 Aaron Sloman, Professor of AI and Cognitive Science at the University of Sussex, cited it as an error that should be avoided in future programmes, in his letter to Sir Austin Bide in 1986: “[this mistake was] forcing people to use totally unsuitable hardware and software just because it is British, thereby holding up significant research and diverting precious highly skilled manpower from advanced research to low-level software development”.619

It was also unclear afterwards what had been gained from involving small firms in the project. Of about 113 firms that participated in the programme, at least 50 were SMEs (according to the European Commission’s definition of SMEs as having less than 400 employees).620 Several of these small firms played invaluable parts in some projects and acquired technology with a relatively small investment. They probably also benefited from the programme’s publicity.621 However, many small firms struggled to find the staff they needed, as well as secure their share of the research funds. The overhead of working on a co-operative project was such that many small firms decided to form partnerships with larger ones and effectively act as sub-contractors.622 The evaluators of the programme indicated that to some extent the overheads were fixed costs and thus they often made collaborative research less attractive for smaller projects and smaller firms.623

Communication between the participants was another challenging area. The programme was not organised around one research centre (unlike the Japanese and American programmes—see below). Instead it had been intentionally decentralised, as distances within the UK were relatively small, academics had teaching responsibilities in their universities, and the relocation of the researchers’ families presented difficulties.624 Moreover, it was considered that a single site would exacerbate the problems of technology transfer back into the firms for exploitation.625

The Alvey programme compared

The Alvey programme was launched in a context of global enthusiasm for large-scale advanced IT research projects. The Japanese Fifth Generation Computer programme, started in 1982, was a ten-year initiative that had the objective of creating a new kind of computer. The term ‘Fifth Generation’ was coined by the Japanese to describe what they thought was going to be the new wave of computer developments. The new computer would use VLSI circuits and new software languages, it would be structured to process information in parallel rather than in sequence, it would exploit a new human-machine interface using speech and image devices and, most importantly, it would be geared towards problem-solving using developments in AI to manage concepts rather than numbers.626

At the end of the ten-year period, the programme had spent over 50 million JPY (the equivalent of £416 million today). This initiative was seen by the European and the American IT sectors as an attempt from the Japanese government to place the country in an advantageous position in the area of computing—as they had done with electronic consumer goods and, to an extent, with the motor vehicle industry.627

The USA saw the Japanese programme as a threat to their global supremacy over the computer and informatics industries.628 Two programmes were subsequently created: the Strategic Computing Initiative (SCI) and the Microelectronics and Computer Technology Corporation (MCC). The SCI was a ten-year programme intended to develop advanced computer hardware and AI. It cost $1 billion USD (the equivalent of £1.85 billion today), provided by the Defense Advanced Research Projects Agency (DARPA). The SCI was conceived as an integrated programme, which meant that different subsystems were created, all working towards the same goal: creating machine intelligence. The MCC was the first American computer industry R&D consortia. The initial budget for MCC’s activities was between $50 and $100 million USD per year (the equivalent of £92 to £184 million today), depending on the number of participants and projects.629

Europe’s response to the Japanese programme was the European Strategic Programme for Research and Development in Information Technology (ESPRIT). The main objective of this initiative was to promote collaboration between European countries in IT R&D. Indeed, every project had to bring together companies and research institutions from at least two EEC countries.630 Moreover, a significant part of the programme was devoted to increase the interaction between users and developers, disseminate results widely, and boost product and process adoption in the market. In the early 1990s, 20% of the overall funding was dedicated to the integration between R&D and take-up.631 The programme of ESPRIT was not fixed from the start: it was adapted every year after extensive consultation with both suppliers and users, in order to reflect the industry’s changing priorities.

There are a number of observations that emerge from the comparison between these projects. First, the Alvey programme was the shortest programme. In the United States, MCC was dissolved in 2000 and the SCI lasted for ten years, like the Japanese Fifth Generation initiative. At its start, ESPRIT had goals that extended over ten years, but subsequent programmes were created and it was only in 1999 that ESPRIT 4 was replaced by the Information Society Technologies (IST) programme.

The Alvey Directorate assumed from early on that there would be a follow-up programme, and in some instances built programmes that were only viable if they lasted ten years.632 The Government and industry, however, saw the Alvey programme as a one-off five-year programme intended to stimulate the community and make the UK an effective international partner or competitor.633 It is not clear why there was such a disparity of views. In 1988 the Government rejected the proposal of a follow-up programme and instead decided to put more emphasis on the UK’s participation in ESPRIT. Indeed the UK began to participate in ESPRIT, on a relatively marginal basis, during the 1980s, but the Alvey programme had absorbed most of the attention and resources during this period. The misunderstanding in terms of the expectations and timeframe of the programme seems to have contributed to disappointment from both parties, as the Directorate was expecting more time to fulfil its goals, while the Government expected results that the programme was never designed to achieve.

The evaluators of the Alvey programme noted that supporting pre-competitive research was a necessary but limited aspect in enhancing the competitiveness and performance of the UK’s IT industry.634 Therefore, they suggested that if improving competitiveness was the objective, governmental and private initiatives had to complement pre-competitive R&D.635 These complementary measures should include:

“more concerted efforts to involve IT users with the scope of R&D programmes, thus stimulating greater user awareness and alerting IT producers at an early stage to the needs of users;

greater effort should be made within firms to formulate technology strategies to facilitate the exploitation of R&D. In some cases it might be appropriate for Government to encourage the process;

a serious re-evaluation of mechanisms to cope with the need for ‘patient’ capital in the further development and exploitation of the enhanced R&D base created by programmes such as Alvey.”636

ESPRIT, and to a lesser extent the American programmes, gave significant support to the interaction between users and developers, the dissemination of results, and the promotion of production and adoption of IT products in the market.637 Drawing from this experience, proposals for an ‘After-Alvey’ follow-up programme suggested a heavier focus on increasing the application of IT and on enabling a closer involvement between suppliers, researchers and users.638 However, with the shift towards European research projects in the late 1980s, this programme was never approved.

Another important lesson that the evaluators and directors of the Alvey programme drew from the comparison between the implementation of the programme and that of ESPRIT was related to the terms of collaboration between members. At the beginning of the programme, the Alvey Directorate had assumed that firms would have had prior experience of drawing up collaborative agreements; since this was not the case, the start of many projects was delayed.639 The evaluators thus suggested that future programmes should emulate ESPRIT’s standard agreements embodied in the contracts, with intellectual property rights held by all participants.640

Finally, a brief comparison of the results of these programmes can be fruitful. The fact that ESPRIT facilitated communication between users and developers granted the programme a great deal of flexibility. The annual revision of ESPRIT’s programme, along with the establishment of industrial advisory panels (which made recommendations from the perspective of IT users), kept the programme in line with the demands of the IT sector.641 Moreover, additional support was provided to firms, and especially SMEs, that wanted to incorporate novel electronics (to them) into existing product lines.642 The diffusion service, PROSOMA, also served an important role in rendering visible the way in which companies within the IT industry had used their participation in ESPRIT to improve their competitiveness in world markets, as well as both publicising and facilitating the use of ESPRIT results by different agents.643

The areas in which ESPRIT results were applied were greatly varied, and included:

Although the SCI did not provide such an open exchange mechanism, DARPA was concerned with investing in technologies which showed the most promise. Therefore, in 1987 its new director Dr Jack Schwartz decided to abandon the focus on AI (e.g. autonomous land vehicles), and to prioritise instead hardware and software with military and civil applications in the shorter term.644 This allowed the SCI to make significant contributions to supercomputing, crucial in weapons design, code breaking, and other areas of national defence.

The goals of the Japanese programme, on the contrary, remained unchanged throughout its implementation. In the 1990s, this lack of flexibility was seen as a weakness by American observers due to the changing nature of the computer industry—which meant that by the time the programme ended, some of its findings were already obsolete or no longer in demand.645 Although the programme did not produce an intelligent machine, it did manage to develop prototype computers that could function at high speeds and the software to control and program them. Overall, the Japanese programme was not considered a success by the international IT community, and indeed the US continued to be the leader in computer design and software in the 1990s.

Therefore, despite having generated a great deal of anxiety, the Japanese programme actually seems to have been the least successful of all the R&D initiatives of the 1980s. Moreover, focusing on the links between research and its application in both the civilian and military sectors appeared to yield good results. Pre-competitive research, although capable of strengthening the IT community, was not sufficient to enhance the competitiveness and performance of the UK’s IT industry.

Conclusion

The idea of ‘AI winters’ in the UK obscures important conclusions that can be drawn from the experience with the Alvey programme. Funding for particular R&D projects does seem to have spiked and dropped at particular points, the latter often occurring after disillusionment had set in. Yet the Alvey programme’s results and implementation problems show that lessons from the 1970s were not properly considered, and that projects were set up without a clear understanding of how commercial and ‘public good’ objectives would be achieved and sustained. This was the case especially in terms of skills, and developing sustainable SMEs. Furthermore, the UK also neglected existing assets, including a highly skilled female workforce.

Professor Lighthill’s assessments, for all of the criticism which has been levelled at them since, were arguably prescient. A general-purpose system was not created in the twentieth century and it is still today not considered a feasible goal by many members of the field. In addition, many of the objectives of AI since the 1960s, such as translation and speech recognition, are only nowadays yielding concrete results, a situation that corresponds more or less with Professor Lighthill’s expectations.

Moreover, Professor Lighthill’s suggestion of promoting the involvement of AI research with the field of application appears to have produced satisfactory results in other countries, and it seems likely that the UK could have benefitted from a similar approach in the 1980s. Indeed, the evaluators of the Alvey programme concluded that pre-competitive research, although capable of expanding and consolidating the IT research community, had not been sufficient to improve the competitiveness of the UK’s IT industry.

Some important points about AI R&D policy emerge. First, the lack of long-term evaluation of Government-backed projects hinders policy planning. Indeed, although an evaluation was carried out shortly after the Alvey programme, there was no follow-up, and it therefore remains unclear what the long-term achievements of the Alvey programme were. Second, to avoid misunderstandings and disappointment, short and long-term objectives should be clear and explicit from the outset between the various participants and funding bodies of R&D projects. And finally, the advice of expert investigators, external to the field (such as Professor Lighthill), appears to have produced useful assessments which could have informed and improved AI policy, had it been heeded. This would suggest that in future, similar evaluations should consider comparable policy developments elsewhere in the world, and past policy successes and failures within the UK. This would enable them to present suggestions which were both well informed and impartial, which could prove invaluable when determining the future goals of AI policy and the best way to achieve them.


586 Science Research Council, Artificial Intelligence: A paper symposium (1973): http://www.chilton-computing.org.uk/inf/literature/reports/lighthill_report/contents.htm [accessed 5 February 2018]

587 Science Policy Research Unit, University of Sussex and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, (Norwich, United Kingdom: Her Majesty’s Stationery Office, 1991), p iv

588 The term ‘pre-competitive research’ was intended to delineate so-called enabling technologies, which on their own did not have viable commercial applications, but were considered necessary for the subsequent development of commercially competitive products and systems by private companies. Brian Oakley and Kenneth Owen, Alvey: Britain’s Strategic Computing Initiative (Cambridge, Massachusetts and London: The MIT Press, 1989), p 3

589 Thinking Machines: The Quest for Artificial Intelligence—and where it’s taking us next, p 3

590 The European Strategic Program on Research in Information Technology (ESPRIT) was a series of integrated programmes of information technology research and development projects and industrial technology transfer measures which ran from 1983 to 1998.

591 The National Archives, Records of the Prime Minister’s Office: Correspondence and Papers, 1979–1997, PREM 19/2116: ‘EDUCATION. New blood for research and information technology: follow-up to the Alvey Report on Advanced Information Technology; international collaboration; the EUREKA programme’, (1982–1987): http://discovery.nationalarchives.gov.uk/details/r/C16204995 [accessed 23 March 2018]

597 Professor Sutherland also recommended improving the connections with the USA given the highly developed state of its AI field. He suggested doing this by facilitating and encouraging the movement of students and researchers between the two countries. Artificial Intelligence: A paper symposium, p 26

598 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p 15

599 For instance, in the 1980s, the Medical Research Council received a number of research grant applications that included ‘Expert Systems’ and referred to them as the application of ‘AI’ to medicine. The National Archives, Medical Research Council: Registered Files, Scientific Matters (S Series), FD 23/2286: ‘Artificial Intelligence Advisory Group: grant applications considered by the Group; notes and correspondence’ (1986): http://discovery.nationalarchives.gov.uk/details/r/C2516889 [accessed 23 March 2018]

600 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p i

601 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p iii

602 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p 141

603 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p 144

604 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p iii

605 The Cambridge advanced IT ‘phenomenon’ was already happening beforehand. Brian Oakley and Kenneth Owen, Alvey: Britain’s Strategic Computing Initiative (Cambridge, Mass and London: The MIT Press, 1989), p 111

606 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p 5

607 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p iv

608 Ibid.

609 Brian Oakley and Kenneth Owen, Alvey: Britain’s Strategic Computing Initiative (Cambridge, Mass and London: The MIT Press, 1989), p 104

610 Alvey: Britain’s Strategic Computing Initiative, p 106

611 This does not refer to the present-day Alan Turing Institute based in London, but to the Turing Institute based in Glasgow between 1983 and 1994. Alvey: Britain’s Strategic Computing Initiative, p 117

612 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p vii

613 Ibid.

614 Marie Hicks, Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing (Cambridge, MA: The MIT Press, 2017), p 208

615 Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing, p 207

616 Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing, pp 214–216

617 Alvey: Britain’s Strategic Computing Initiative, p 117

618 Alvey: Britain’s Strategic Computing Initiative, p 109

619 A. Sloman, Letter to Sir Austin Bide, (undated) 1986, cited in: Alvey: Britain’s Strategic Computing Initiative, p 158

620 Alvey: Britain’s Strategic Computing Initiative, p 111

621 Ibid.

622 Ibid.

623 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p iii

624 Alvey: Britain’s Strategic Computing Initiative, p 113

625 Ibid.

626 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p 5

627 Ibid.

628 Japan had been accused by members of the American IT industry of profiting from other countries’ innovations, which they considered unfair. Copying (as opposed to inventing and innovating) was not considered a threat. Andrew Pollack, ‘’Fifth Generation’ became Japan’s Lost Generation’, New York Times (5 June 1992): https://www.nytimes.com/1992/06/05/business/fifth-generation-became-japan-s-lost-generation.html [accessed 23 March 2018]

629 Michael Frontain, ‘Microelectronics and Computer Tech Corporation (MCC)’, Handbook of Texas Online (15 June 2010): https://tshaonline.org/handbook/online/articles/dnm01 [accessed 6 September 2017]

630 European Commission, Community Research and Development Information Service (CORDIS): www.cordis.europa.eu/esprit/src/intro.htm [accessed 6 September 2017]

631 Ibid.

632 John Fairclough (Government’s Chief Scientific Adviser) cited in: Alvey: Britain’s Strategic Computing Initiative, p 259

633 Alvey: Britain’s Strategic Computing Initiative, p 260

634 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p vii

635 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p viii

636 Ibid.

637 The American programmes also granted considerable attention to users’ needs.

638 Alvey: Britain’s Strategic Computing Initiative, p 229

639 Alvey: Britain’s Strategic Computing Initiative, pp 109–110

640 Evaluation of the Alvey Programme for Advanced Information Technology: A Report by Science Policy Research Unit, University of Sussex, and Programme of Policy Research in Engineering, Science and Technology, University of Manchester, p xi

641 European Commission, Community Research and Development Information Service (CORDIS), ‘Archives’: www.cordis.europa.eu/esprit/src/intro.htm [accessed 8 March 2018]

642 This service was called FUSE (First Users).

643 European Commission, Community Research and Development Information Service (CORDIS), ‘Case studies from ESPIRIT—Information technology for business’: http://cordis.europa.eu/news/rcn/7343_en.html [accessed 8 March 2018]

644 Paul Ceruzzi, ‘Strategic computing: DARPA and the Quest for Machine Intelligence, 1983–1993 (review)’ in The Journal of Military History, vol. 67, no. 3 (July 2003), pp 994–996:

645 Andrew Pollack, ‘’Fifth Generation’ Became Japan’s Lost Generation’, The New York Times (5 June 1992): http://www.nytimes.com/1992/06/05/business/fifth-generation-became-japan-s-lost-generation.html [accessed 8 March 2018]




© Parliamentary copyright 2018