Digital exclusion Contents

Chapter 8: Accessible services

User-centred digital services

161.In our inquiry we explored why the shift towards digital was typically viewed as a ‘good thing’ by service providers when it left many customers frustrated, reduced in-person interactions and rendered sections of the population unable to use valued services.274 The House of Lords Covid-19 Committee found many areas where digital has proven to be “a very poor substitute for ‘in person’ services and interactions”.275

162.Joanna Causon, CEO of the Institute of Customer Service, said most users preferred digital interaction but “a small but significant minority, around 15 per cent” struggled.276 She emphasised that essential services should maintain offline alternatives.277 Owen Barry, Managing Director of Justice, Central Government and Transport at Capita, thought that digital transformation efficiency gains had stagnated somewhat in recent years but believed that such projects were “driven by customer and general public expectation” rather than pure cost-saving.278

163.We heard that the accessibility of digitised services, both in their design and associated support provided, often did not meet expectation or demand. The Government’s 2012 ‘Digital By Default’ strategy for public service delivery was a good example. It aimed to make central government smaller, faster, more unified, more accountable and more commercially capable.279 The Government said digital services would be “so straightforward and convenient that all those who can use digital services will choose to do so, whilst those who can’t are not excluded and that those in the latter category would be supported through the principle of ‘Assisted Digital’.”280

164.Yet we heard that the consequent shift to digitised services has intensified the social and economic exclusion among many groups.281 Universal Credit was cited. A 2020 National Audit Office review found that only around 20 per cent of Universal Credit applicants were able to verify their identity online, and highlighted concerns that people with low digital skills might find it particularly difficult to provide the evidence required and submit claim applications.282 Patricia Bailey from the poverty charity APLE Collective said it had created “too many hurdles where, if you cannot provide an email address, you cannot claim.”283

165.Paul Waller, Research Principal at Thorney Isle Research, said that “central government has from time to time proposed to provide ‘offline access to services’ … but this has not always materialised.”284

166.Professor Simeon Yates told us:

“a lot of digital solutions push the problem down … we are still treating the digital delivery of public services a bit like Ryanair. ‘We’ll make a huge saving. We’ll push the process of administering whatever this social service is on to the client’.”285

167.We heard that countries such as Iceland had prioritised local, in-person support to accompany their digital transformation strategies. Róbert Bjarnason, President of Citizens Foundation Iceland, explained that resources were targeted to libraries, city service centres and closed-down bank branches to transform them into digital support centres.286 Kristina Reinsalu, Programme Director of e-Democracy at the e-Governance Academy Estonia, said Estonia had focused its limited resources on public libraries “in even the smallest rural areas and villages, where we launched free-access internet points with mentors and people who could [provide] support”.287

168.The Government said it was already supporting in-person hubs, for example by providing wi-fi in libraries and support for community organisations. We noted however that the current level of support suggests staff in these organisations may lack sufficient resources and training to address the scale and variety of help needed. Libraries Connected, a representative organisation, said that while libraries were playing an increasingly important digital inclusion role, staff “cannot meet the demand” due to a lack of capacity, restricted opening hours and limited equipment.288 The Local Government Association also told us that pressure on council budgets has led to a 43.5 per cent net decrease in expenditure on libraries between 2009 and 2019.289

169.The shift towards digital by default public services has not been accompanied by commensurate support for those who struggle with digital access. Libraries and community organisations have taken on additional responsibilities to fill these gaps, but without sufficient resources and training.

170.The Government’s digital inclusion strategy refresh should include support for place-based in-person initiatives to help those who cannot navigate online access to essential services. This could include boosting the role of libraries, community centres and local amenities as inclusion hubs, in partnership with businesses.

171.Not everyone wants to be online, or online all the time. And some services are better in person. Private and public service providers should avoid viewing digital as a cheap substitute for good customer service. Adequate provision must be maintained for those who cannot or do not wish to use online services.

Accessible design

172.Making online service designs more accessible would help address a significant obstacle to digital inclusion for people with disabilities. Some assistive technologies are available, such as screen readers for those with visual impairments, though they have limitations and not everyone can use or afford them.290 In 2018 the Government issued regulations for public sector websites to improve accessibility. By 2020, some 74 per cent of UK council websites still lacked adequate features for screen readers, however.291

173.The British Academy thought more attention should be paid to the Government’s Service Standard manual, which sets digital service provision and inclusive design standards.292 It argued there was “little pressure on departments to uphold this standard or for the National Audit Office to audit it.”293

174.Dr Robin Christopherson, Head of Inclusion at AbilityNet, thought existing regulations provided a good starting point, noting the Equality and Human Rights Commission and the Cabinet Office monitored compliance and could impose fines.294 He suggested further improvements could be achieved by encouraging businesses to see the commercial benefit of providing inclusive services; and by extending public sector standards to private sector services.295

175.We also noted that improving accessibility standards for people with disabilities would further enable them to use digital services with independence and privacy. This is particularly significant when accessing health and other services which require entering sensitive personal information. A recent report found that enabling those with disabilities to use digital services with confidence “might lessen the pressure on support staff and give people with learning disabilities more control and agency.”296 Evidence from APLE Collective, a network of individuals who experience poverty, highlights that being forced to rely on support from others to access digital services, particularly in public places like libraries, can compromise privacy.297

176.Too many online services have poor accessibility for those with additional needs. As part of the strategy refresh, the Government should audit public sector websites for compliance with accessibility standards and regulations.

177.The Government should encourage private sector organisations to adopt website design accessibility standards used by the public sector. This could start with the most significant public-facing services, for example in healthcare, finance and housing.

Predictive analytics

178.We heard that digitally excluded groups may be poorly served by trends towards greater use of machine learning and predictive analytics in public-facing services. The Alan Turing Institute notes that governments “are major holders of data which data science and AI can harness to improve the design and provision of public services.”298 The use of machine learning tools in decision-making is growing across Government and public services, according to the Centre for Data Ethics and Innovation.299

179.Most tools are still in the initial phase of development and deployment.300 Professor Helen Margetts, Director of the Public Policy Programme at the Alan Turing Institute, told us that local government has “been more keen and probably used these technologies faster than central government.”301 Dr Adrian Weller, Director of Research in Machine Learning at the University of Cambridge, said that in 2018 around 53 out of 96 local authorities surveyed and about a quarter of police authorities were using algorithms for prediction, risk assessment and assistance in decision-making.302

180.We heard that digitally excluded groups were likely to be underrepresented in some data sources, whilst belonging to demographics that are typically overrepresented in other sources.303 Professor Helen Margetts explained that this could have practical consequences across a range of areas, particularly if the data are used to inform service improvements or resource allocation decisions. She gave the example of trying to complain about a poor service:

“In the end, the only thing that got me noticed was going on to Twitter … which [the providers] were actually recommending … So think about their complaints data. Somebody who was digitally excluded could not complain during that period; they had no way to complain. They are not in the data so, if they run any sort of algorithm on their complaints data, those people are not there, and they do not know what their problems are.”304

181.We noted that there did not appear to be an overall assessment on the use of such tools across central and local public services and the implications for digital exclusion policy.305 Rachel Coldicutt, Executive Director of Careful Industries and Promising Trouble, noted that automated decision-making in welfare policy was already “common throughout Europe” and thought it was likely to become a feature across multiple sectors including justice, migration and health. She cautioned about the importance of safeguards, highlighting the example of the Netherlands, where faulty data led to 20,000 families being wrongly accused of child benefit fraud in 2021.306

182.As public-facing services become increasingly digitised, machine learning tools and predictive analytics are likely to influence policy choices and service delivery. Digitally excluded groups are at risk of being poorly represented in key datasets, and hence face further marginalisation.

183.The Government should commission a review to understand the extent of predictive analytics in public-facing services, their likely trajectory over the next five years and the effects on digital exclusion policy. In the meantime the Government should require public service providers using predictive analytics to consider the use of data and the impacts on those who are digitally excluded.

275 Covid-19 Committee, Beyond Digital: Planning for a Hybrid World (1st report, Session 2019–21, HL Paper 263), p 3

281 Q 86 (Professor Simeon Yates); Digital Poverty Alliance, UK Digital Poverty Evidence Review 2022 (June 2022) p 9: [accessed 16 May 2023]

282 National Audit Office, Universal Credit: getting to first payment (Session 2019–2021 , HC 376), p 12: [accessed 16 May 2023]

284 Written evidence from Paul Waller (DCL0010)

288 Written evidence from Libraries Connected (DCL0033)

289 Written evidence from the Local Government Association (DCL0062)

290 Written evidence from National Network of Parent Carer Forums (DCL0059)

291 Socitm, Top five most common accessibility issues faced by UK council websites (2020): [accessed 16 May 2023]

292 HM Government, ‘Make sure everyone can use the service’: [accessed 15 May 2023]

293 Written evidence from the British Academy (DCL0023)

296 Magdalena Mikulak, Sara Ryan, Siabhainn Russell, Sue Caton, Richard Keagan-Bull, Rebecca Spalding, Francesca Ribenfors, Christopher Hatton, ‘Internet is easy if you know how to use it: Doing online research with people with learning disabilities during the COVID-19 pandemic’, British Journal of Learning Disabilities, vol. 51, issue 2 (2022), pp 269–278:

297 Written evidence from APLE Collective (DCL0013)

298 The Alan Turing Institute, ‘Public Policy’: [accessed 15 May 2023]

299 Centre for Data Ethics and Innovation, Review into bias in algorithmic decision-making (27 November 2020): [accessed 16 May 2023]

300 Centre for Data Ethics and Innovation, Review into bias in algorithmic decision-making (27 November 2020): [accessed 16 May 2023]

305 Q 70 (Professor Helen Margetts)

306 Written evidence from Careful Industries and Promising Trouble (DCL0088)

© Parliamentary copyright 2023