At Apolitical, we’re excited by the transformative potential of Artificial Intelligence (AI). Our mission is to help build 21st-century governments that work for people and the planet, and we believe AI can help us achieve it. We also believe that AI presents governments across the world with unique opportunities to improve the lives of citizens and the environment, and we believe our role is to help these impacts be realised. You can find more information about Apolitical’s approach to AI in our AI Charter
But, as with all new technologies, AI requires strong governance to ensure it is used responsibly, fairly and sustainably. Therefore, we have developed a set of AI Principles that form the basis for our team about our expectations for use of AI internally, and an Acceptable Use Policy that applies to any use of our externally facing AI-powered tools. These principles are founded in global best practices and frameworks governing the safe use of AI for governments, including the EU AI Act, and guidance from the Information Commissioner’s Office.
Transparency
To ensure that our members are comfortable engaging with AI systems on the Apolitical platform, they need to know where and how they are used. As such, in line with our Terms and Conditions, we will endeavour to highlight where AI is used. We also will share what data our AI models are trained on and provide a high-level explanation of how it works. This information can be found in the table below, which is subject to change as we develop our platform and products.
Used for Good
We use AI to drive us towards achieving our mission. As such, we only use AI where it positively impacts our members, the governments they operate in and society more generally. We embed this principle into decision making and product design for AI-powered tools, and ensure we have stress tested potential impacts through developed governance processes.
Fairness
We design and use our AI systems to the equal benefit of all our members and, beyond that, all groups in society for which our members work. We are constantly improving our AI systems to ensure that biases are mitigated. We also carefully consider the impacts of AI systems on society and if there are any disproportionate impacts on a specific group before we deploy it.
Safety and Security
We build and use AI systems with the principle of privacy by design and default, a core consideration at each stage of the AI lifecycle. We also ensure our team is appropriately trained to engage with and develop AI systems. We only use AI systems that ensure transparency, security and control. We also minimise the data input into AI systems, and don’t process sensitive data or personally identifiable information using AI systems without giving you clear prior notice, in line with our Privacy Policy.
Accountability
We see AI systems as ways to enhance and augment human work, not replace it. We therefore ensure that human oversight is required for any decisions or workflows that legally, contractually, or ethically must be reviewed by a human. Our internal policies place clear accountability on staff, not AI, for judgements, decisions made and actions taken with the use of AI as part of our work. We recognise that Apolitical is responsible for the deployment and monitoring of AI systems on our platform. As such, we pay close attention to the feedback we receive, and constantly iterate on and improve the models we deploy. We also implement processes for monitoring for inappropriate or unlawful use of AI systems on our Platform.
AI Acceptable Use Policy
As part of our commitment to safe, responsible and innovative AI development and use, we set out our expectations for users of AI tools available on the Apolitical Platform in this Acceptable Use Policy. This Policy sets out the expectations of use for all individuals who engage with AI-powered tools on Apolitical. This policy should be read in conjunction with our Terms and Conditions and Privacy Policy.
Usage Policy
In line with the principles of Used for Good and Safety and Security set out above, we have created a set of Usage Policies which outline unacceptable and prohibited uses of AI tools which are available on the Apolitical system. In using such AI tools, you are responsible for ensuring your use complies with this policy.
We conduct monitoring for unacceptable, illegal or inappropriate uses of AI across our platform, and may at our discretion revoke your access to the Platform and our tools in response to a potential or real breach of these terms, in line with our Terms and Conditions.
You are responsible for ensuring the accuracy of the information you input into AI tools, and for ensuring that you do not use AI-powered tools on the Apolitical Platform for any of the following purposes:
Illegal Activities
You may not use any service on the Apolitical platform for an activity that is illegal under the laws of England and Wales, or any laws of the jurisdiction you are subject to. This includes, but is not limited to:
- Generating, disseminating or facilitating content that constitutes or incites hate speech, discrimination or harassment on the grounds of race, ethnicity, nationality, religion, gender, sexual orientation, disability or any other protected characteristic under the UK’s Equality Act 2010 or equivalent legislation.
- Creating, distributing or facilitating any content that is defamatory, fraudulent, deceptive or constitutes misrepresentation.
- Any activity that violates intellectual property rights, including copyright infringement or the unauthorised reproduction of third-party content.
- Any activity involving the unlawful collection, processing or transfer of personal data, including data scraping or profiling without lawful basis.
Any activity that facilitates bribery, corruption, money laundering, terrorist financing or sanctions evasion, contrary to the Bribery Act 2010, Proceeds of Crime Act 2002, or equivalent legislation.
Harmful Activities
- Generating or distributing content designed to manipulate, deceive or psychologically harm individuals or groups, including the creation of disinformation or influence operations.
- Generating or distributing content that targets, exploits or endangers children or vulnerable persons.
- Generating or distributing content that promotes or facilitates self-harm, suicide or violence against any person.
- Attempting to use AI tools to circumvent, manipulate or undermine democratic processes, electoral integrity or governmental decision-making, including lobbying or launching personalised campaigns.
- Generating or distributing deepfakes, synthetic media or other content that falsely attributes statements or actions to real individuals, particularly public officials or government representatives.
- Any activity that could cause material harm to Apolitical's reputation, its members, partners, or the communities it serves in line with our Values and Guidelines
Security
- Attempting to reverse-engineer, extract, replicate or otherwise misuse the underlying models, training data, prompts or infrastructure of any AI tool on the platform.
- Using AI tools to conduct, plan or facilitate cyber attacks, social engineering, phishing, malware creation or any other malicious technical activity.
- Attempting to manipulate AI tools through prompt injection, jailbreaking or other adversarial techniques intended to circumvent this Policy or Apolitical's Terms and Conditions.
- Using AI outputs to impersonate Apolitical, its employees, partners or any other organisation or individual.
Misuse of the platform’s purpose
- Using AI tools for purposes unrelated to professional development, learning, policy knowledge or the legitimate public sector objectives for which the platform is designed.
- Using AI tools to generate content for commercial gain outside of the scope of your access agreement with Apolitical, or in breach of any institutional licence.
- Using AI tools to generate or circulate political propaganda, partisan advocacy or campaign materials.
Personal, classified, and sensitive data
- You may not enter the personal data, including special categories of data as defined in Article 9 of the UK GDPR, of other individuals into any AI tool unless you have independently verified that you may do so in compliance with all applicable data protection laws.
- Do not enter data that is classified, restricted or marked as confidential by your government, organisation or institution
- Do not enter trade secrets, commercially sensitive information or proprietary data belonging to your organisation or a third party without appropriate, written authorisation
Use of Outputs
In line with our AI Principles, Apolitical believes AI should enhance human work, not replace it. We expect users to apply the same principle in their own engagement with our tools, and ensure that you apply appropriate judgement and verification when using outputs.
You must not use AI-generated outputs without adequately fact checking or verifying the accuracy of the outputs. You must also ensure that decisions that carry material consequences, including policy decisions, legal determinations, resource allocation, personnel decisions or any action that could affect the rights, wellbeing or interests of individuals or communities are appropriately reviewed by a qualified human, or group of humans.
You remain subject to your organisation's own policies, codes of conduct and ethical obligations. This Policy does not limit or replace those obligations; it operates alongside them. You should consult your institution's guidance on the use of AI tools where relevant.
As stated in our Terms and Conditions, the Apolitical platform and content is provided for informational purposes only, and should not be considered to constitute advice. This applies to all AI-powered tools provided as part of the Apolitical platform. Apolitical does not warrant the accuracy, completeness or quality of the outputs provided by AI tools available on the platform.