Dear colleagues,
Generative AI (ChatGPT, Dall-E, Bard, even DeepL i.a.) is a powerful tool that's rapidly evolving. It can help us in numerous ways, from creating communication resources for instance, to assisting in technical projects. However, with great power comes great responsibility.
So much power in fact, that the safe development and use of Generative AI is being defined on a global level. Earlier this month, during the AI Safety Summit held in Bletchley Park (UK), the Bletchley Declaration on AI Safety was signed by 27 countries. This declaration, a clear statement of the international community’s commitment to ensuring that AI benefits all of humanity, outlined a set of principles, including the development and use in such a way that it benefits all of humanity, respects human rights and fundamental freedoms and allows for a transparent and accountable use.
Following this international example of affirming that AI should be used in a manner that is safe, human-centric, trustworthy and responsible, expert colleagues at BESIX Group are currently working on a clear Group Policy on how to make the most of this technology while ensuring we use it safely and ethically for professional purposes.
In the meantime, we want to address the following for those among you already experimenting with AI: do it responsibly.
While we encourage you to explore the capabilities of generative AI, always be aware of the potential risks and outcomes.
The attached guidelines are a first step to accompany you in this discovery. Please read them carefully.
Make sure to always keep these 5 rules in mind when using generative AI.
- Don’t use confidential information in generating AI prompts: always remove any confidential or personal information from your inputs before using generative AI.
- Don't rely solely on AI: while AI can provide quick solutions, it's essential to review its outputs. It's not infallible. Always fact-check the output.
- Steer clear of sensitive subjects: generative AI might not always handle delicate topics with the nuance they require.
- Avoid plagiarism: generative AI can sometimes produce content that resembles existing work. Always ensure the content is original and doesn't infringe on any copyrights.
- Don't use AI blindly: if you're using AI to generate content in a language you're not familiar with, always consult with a native speaker.
Going forward, you can expect a policy and gradually, the diffusion of examples of user cases for specific usage (ex. for prequal, tender, communication, ESG…), and potentially the creation of a BESIX-centered AI platform.
E-learning and various resources will be created as well, and distributed as we grow as a group in the mastery of these new tools.
Remember, while generative AI can be a fantastic assistant, it's just a tool. We're still responsible for the final output. Let's use it wisely, ethically, and responsibly.
Stay curious and safe, while experimenting with AI!
BESIX Group ICT Department