The generative artificial intelligence firm whose tech is the engine behind GOV.UK Chat was recently found to have accessed millions of pirated works, resulting in a $1.5bn settlement with authors
Government is facing questions about the use of pirated intellectual property by the artificial intelligence firm whose systems underpin the GOV.UK Chat service, as well as the environmental credentials of its tech partner.
When the first trials of the government chatbot began in 2023, the tool was built on technology from OpenAI – the firm behind large language model ChatGPT. But more recent versions of government program have instead used a rival generative platform: the Claude LLM from Anthropic.
Following a lengthy court battle, a judge’s ruling earlier this year found that Anthropic had knowingly accessed digitised copies of seven million books that had been pirated. A lawsuit led by a group of authors subsequently resulted in an agreement by the AI firm to pay out a $1.5bn settlement.
In light of the findings of this case, Liberal Democrat peer Tim Clement-Jones filed a series of parliamentary questions enquiring as to whether the Government Digital Service, which developed GOV.UK Chat, was aware of the findings that emerged from the case, and “whether they conducted a risk-assessment for the use of a model developed using unlawfully accessed copyright-protected works”, as well as considering wider “ethical and legal considerations” of working with the California-headquartered firm. He also asked if GDS had instead considered working with UK-based specialists.
In response, digital economy minister Baroness Liz Lloyd said: “The Government Digital Service recognises the importance of ethical and legal considerations relating to the use of copyright-protected material in the training of large language models.”
The minister suggested that GOV.UK Chat’s current use of Anthropic’s technology is a “phase” in which the Claude LLM is being assessed – and that the government chatbot may incorporate other generative tools in future phases.
“A range of large language models from multiple suppliers have been considered in developing GOV.UK Chat,” Lloyd said “Earlier iterations tested OpenAI’s ChatGPT models, while the current phase is evaluating Anthropic’s Claude models.”
She added: “GOV.UK Chat is a product in active development that currently accesses Anthropic models through an existing agreement with Amazon Web Services, enabling the Government Digital Service to test a range of models from different suppliers. Risk assessments have been undertaken in accordance with government standards, including consideration of ethical, legal, and data protection risks relevant to the models tested. GOV.UK Chat is being developed entirely within the Government Digital Service by a multidisciplinary team of civil servants. The Government continues to engage with UK-based AI developers and remains open to collaboration where this supports innovation and the delivery of public services.”
Related content
- Government signs collaboration agreement with ChatGPT-maker OpenAI
- AI firms must address hallucinations before GOV.UK chatbot can roll out, digital chief claims
- EXCL: DWP reverses policy prohibiting staff use of ChatGPT and other LLMs – except DeepSeek
Clement-Jones also pressed government on “what assessment they have made of the environmental and sustainability impact, including energy demands, of the Anthropic Claude AI model used for the GOV.UK Chat service”.
In response, Lloyd indicated that the decision to use Anthropic’s technology already represents the choice of a “more efficient model”, when compared with alternatives.
“The Government Digital Service recognises the importance of environmental considerations in the use of AI and we are committed to using this technology responsibly,” she said. “We are working in line with the AI Playbook for the UK Government, which highlights the importance of understanding and managing the environmental impact of AI systems. Environmental considerations have been reviewed under the Environmental Principles Policy Statement (EPPS), with negligible anticipated impact. EPPS principles have been considered, including resource efficiency and sustainability of digital infrastructure. Model selection decisions are based on performance, security, data protection, cost, and alignment with government standards. Where possible, we use smaller and more efficient models to improve sustainability, including in our use of the Claude models within GOV.UK Chat.”
Recently published departmental records revealed that government believes its new chatbot tool is “beating industry standards” for the accuracy of its response to users.
The technology remains in private beta phase, and 2,000 citizens are currently engaged in a four-week trial, during which the chat service is being made available via the GOV.UK App – where it is intended that the tool will ultimately be fully embedded.
