I hate the be the one campaigning for yet another policy, but her me out! In an era where AI can do so much, we need to commentate to clients how we use these tools. At the end of the day, they are paying for your expertise, not your ChatGPT skills. We also need to provide clear rules for our colleagues and employees regarding their reliance on AI tools and the potential for compromising privacy of data. I’m not saying there is no merit in letting AI tools do the heavy lifting of editing a presentation or finding the information you need (shoutout to myESRA).. All I’m saying is, it is probably time to set the rules around AI use in your company and communicate it to your clients.
This template is a blueprint that you can use to structure your documents. It touches on the important aspects that are relevant to E&S performance management and development finance in general.
What to keep in mind when designing your company’s AI use policy?
Understanding the role of AI in your company
What are the potential needs that could be filled with AI tools?
What applications are you currently using?
What tasks or jobs you could automate?
The in-dept understanding of potential use cases and how they interfere with your project work can help you clarify a few things:
Current and potential use cases
AI tools in use or type of AI tools that have potential use for your organisation
Disclosure if client deliverables involve the use of AI
It is important to flag the need to use AI for efficiency and to increase your competitive edge maybe? It is undoubtable that your company can have a much greater impact though faster processing of development projects. Also important to note that speeding up certain processes around data collection and research can cut costs for clients and enable experts to focus on analysis and designing mitigation measures.
Ensuring data privacy and confidentiality
Does your organisation have a privacy policy or data protection policy specific to how you treat the data of project affected people? I bet they have one for employee data, but not necessarily guidance on protecting sensitive, personal information related to projects.
How well do you know the privacy policies of the AI tools that you use? Especially, for information that is confidential or not disclosed yet?
This is my new favourite topic so I can go on and on about it. The fact is, we all have sensitive PAP-related data on our laptops from all over the world that we most likely store without compliance with GDPR or other similar legislation. Most organisations established stringent policies for the protection of employee and prospective employee data, but they have no clear guidelines on how to store, access and dispose of sensitive project information.
I have written about Copilot before and how problematic it might be to ensure confidentiality and privacy within an organisation through a complicated set of profile access management steps for each employee. I have also written about the fact that most AI applications will use your data to train their algorithm so they read, store and remember your data. That’s why I’m releasing privatESRA in a few weeks!
Setting out your data protection and privacy commitments in your AI Use Policy - together with your data protection impact assessments, will definitely give clients the peace of mind that their confidential documents are not fed to ChatGPT for analysis.
Adhering to legal ethics
While some of you work in the EU - aka the world’s most regulated AI community, others operate in more liberal jurisdictions. Regardless of what is legally required, you might want to raise the bar in terms of ethics and transparency to include the summary of your values and approaches related to the use of AI tools within your organisation.
Implementing training programs
Sometimes people get disappointed in technology because they don’t know how to use it properly. Over the past 10 months - since we launched myESRA, I have trained many people on the use of generative AI tools and it has been life changing for them. They previously thought myESRA or ChatGPT was ‘stupid’ and ‘useless’, but improving prompts, understanding use cases and limitations completely transformed their relationship with these AI systems. Empowering users through knowledge and training them on the types of AI, specific use cases of AI systems the organisation uses and how to interpret the results are essential. We don’t want our people overdependent on AI predictions, especially if they don’t understand how the AI system came to that conclusion..
Keeping up with the times is challenging as every month we are leapfrogging towards artificial super intelligence. In my experience, the E&S space is a much more careful and slower adapter to integrate AI in the workflow. Nevertheless, I have faith that we can design products that will bring the AI revolution into the heart of E&S!
The template is downloadable as a word file below or you can just copy this text:
Template AI Use Policy for your company
Purpose
To establish a comprehensive framework for the ethical, responsible and effective use of artificial intelligence (AI) in COMPANY NAME, aligning with our commitment to client service, data privacy, confidentiality and ethics.
Scope
Applies to all employees, contractors and subcontractors and AI systems used within the firm, as well as applications of external AI systems and services.
AI Utilization and Application
AI will be integrated for tasks such as research, document review, client communication with specific focus on enhancing efficiency and accuracy.
Consider: Evaluate AI toolls for their ability to handle complex tasks without compromising the need for human judgement, especially in areas like data analysis and drafting documents.
Data Privacy and Confidentiality
AI systems must comply with the highest standards of data protection and confidentiality
Consider: Choose AI tools that adhere to a strict policy of not using our data for training their models, ensuring that sensitive client information remains protected and is not utilised for any other purpose.
Ethical use of AI:
AI will be used in a manner that aligns with our ethical obligations, ensuring unbiased and fair recommendations and decisions.
Consider: Regularly review AI tools for potential biases, particularly in areas that cover generative and predictive AI where transparency on how the algorithm makes decisions is limited, ensuring that they are aware of both the capabilities and limitations of the AI, including its tendency to ‘hallucinate’ or provide creative interpretations.
Transparency and client consent
Clients will be informed about the use of AI on their projects, maintaining transparency and obtaining consent where necessary
Consider: Clearly communicate to clients the extent of AI’s involvement on their project research, ensuring that they are aware of both the capabilities and limitations of the AI, including its tendency to ‘hallucinate’ or provide creative interpretations.
Training and education
Comprehensive training will be provided to all personnel on the responsible use of AI tools, emphasising their strengths and limitations.
Consider: Develop training programs that not only educate about AI capabilities but also encourage staff to apply their judgement in conjunction with AI insights, particularly in interpreting AI-generated predictions and research.
Compliance and monitoring
Continuous monitoring of AI tools to ensure compliance with performance standards and ethical guidelines
Consider: Implement a system for ongoing audits and performance reviews of AI tools, with a focus on ensuring that any evolving algorithms or updates continue to meet our firm’s ethical and professional standards.
Review and Amendments
This policy will be reviewed annually and updated as necessary to reflect technological advancements and changes in legal standards
Consider: Stay abreast of changes in AI technology and legal regulations, ensuring that our policy remains relevant and effective
Implementation
Effective date:
Responsibility for implementation and enforcement of this policies lies with – IT/ Committee/Experts
Comments