How can government build confidence in the cloud?

Public cloud environments was the topic of discussion at a recent roundtable organised by Government Transformation Magazine and supported by Hewlett Packard Enterprise.

Civil servants came from across Whitehall to share their perspectives on how citizen trust could co-exist with the transparency issues presented by a multi-cloud strategy, and AI. The discussion was moderated by our co-founder and former Government CIO, David Wilde.

Rebuilding citizen trust 

Attendees felt that trust between citizens and government had been eroded. The Post Office Horizon IT inquiry and data breaches during COVID lockdown have tarnished citizens’ relationship with the government as reliable custodians of data. As one civil servant put it, "Citizens are more willing to hand their data to social media companies as opposed to an amorphous government".

Though cloud platforms were recognised as having clear utility in the public sector, concerns were expressed - including how data would be stored after a contract ceased: “How we can protect our data?” asked one participant. Cloud service providers need to make the life cycle of public sector data transparent to ensure that government and citizen data is protected. This was highlighted as a source of anxiety with potential leaks fracturing the relationship between government and citizens. One takeaway from this was to learn from the private sector by adopting security measures such as multi-factor authentication and a multi cloud environment, helping to minimise data breaches whilst maximising flexibility for diverse business needs. 

Two of the named solutions to the fractured trust between government and citizens were accountability and transparency.

Accountability was recognised by all attendees as a crucial requisite for AI adoption in government, but there is still fear of the unknown. One civil servant remarked, “If I went to the doctor and they had an AI assistant that was used to support my diagnosis or prescribe me medicine, that would make me extremely uncomfortable”. Despite this uncertainty, there was agreement amongst the attendees that optimism was justified given the way AI is being treated with responsible caution by regulatory bodies.

Transparency was also identified as a major component to building trust. As one civil servant observed, “often when I go to a website, there will be a pop up stating I can’t see what’s on the page without allowing them to use cookies. If I don’t know the content of the site I’m about to explore, how do I know whether to allow third-parties to track me is a fair transaction?”.

Government needs to learn from errors such as these by clarifying exactly how they will benefit from the custodianship of their data to ensure that users can make an informed consent opt-in as you would expect with any other fair transaction.

Multi-cloud environment 

One attendee underlined, “We want to understand the cost perspective and experiment with different leading hosting providers”, whilst another noted, “We’re restructuring and looking at the future of a multi-cloud environment”.

Despite the named benefits, two considerations of multi-cloud environments were brought to the fore.

1. Cost. As AI governance and implementation strategies are increasingly pushed in government, one attendee underlined that they wanted to look beyond AWS and experiment with which hosting providers would be best at driving costs down for integrating Large Language Models into their department’s current digital infrastructure whilst preserving security needs. 

2. Lack of integration. Secondly, there was broad consensus that despite the multi-cloud environment being the norm in government, there was a lack of integration from a lack of standardisation, poor networking capabilities and the reluctance to pool cross-departmental data in case of non-compliance with segregated data protection standards. Incompatible and variable,  government interface protocols hinder data migration between government departments and organisations which perpetuate slow decision-making in government.

Cloud as net zero stepping stone

Two of the most significant forces of change were raised in the roundtable: Net Zero and AI. One attendee stated, “Departments are put under increasing pressure to clarify their net zero plans and this extends to our software too. How is HPE approaching this public sector need?”.

Addressing this, HPE responded that they were “hosting sustainable data centres to promote energy efficiency and will be net zero by 2040”, a timeline “unique” in the market and 10 years ahead of UK government's targets.

Acknowledging that multi-cloud environments are the most attractive strategy for leveraging AI in government, but that poor integration persists in government, HPE highlighted that their new acquisition of networking company Juniper will enable them to meet the public sector’s demand for secure, integrated technology solutions that connect, safeguard, and analyse data across their edge to cloud environments.