# Swiss AI SME Circle #2

In today's group discussion at the second [Swiss AI SME Circle](https://www.swissict.ch/event/einladung-invitation-2-swiss-ai-sme-circle/) hosted by SwissICT, I discussed and finally presented the question "Finance, Legal, Governance, Policy"

The following input was generated using Apertus on the Public AI inference utility based on my notes. I've corrected it to better match the statements. After logging in you can see the full thread here:

https://chat.publicai.co/s/2535febe-3e73-4c20-b62b-ea6d8b98ced7


---

### Summary

Open frontier models like Apertus can help with the use case of supply chain risk in global finance by helping to unpack surveillance capabilities and data protection regulations. We considered the use of satellite data to investigate risk factors, such as ecological developments in supply chain relevant geographies. This is being actively researched by one contributor to our group.

### Use Case

- **Can AI be used to ensure that using satellite data doesn’t inadvertently expose sensitive information or violate privacy/data protection regulations?**
- Satellite data is used to monitor global supply chains, identify potential risks (e.g., natural disasters, infrastructure disruptions, production delays), and provide early warnings to financial institutions and stakeholders.
- This is a good example of end-to-end application of AI technology: from the initial measurements and data collection, to evaluation and recommendations.

### Hurdle

- Unpacking surveillance capabilities, such as explaining what privacy issues exist at particular resolution and processing characterists.
- Uncertainty about the legal and ethical use of satellite data for commercial purposes, especially concerning sensitive data like trade flows and infrastructure locations.
- Lack of clarity on data protection and surveillance regulations in different regions (e.g., GDPR, Swiss DSG, trade secrets).

### Solution/Hack

- Utilizing open Frontier AI Models for scenario comparison.
- Even without multimodal capabilities, Apertus can be used to support the legal work, technical reporting, and more.
- We can use secure and privacy-preserving features of LLMs to anonymize and aggregate datasets, protecting sensitive information and metadata.
- Leveraging ALPS infrastructure could help to host and process data in a compliant and secure environment, adhering to Swiss and global data protection regulations.
- Use Apertus to create a knowledge graph defining sensitive infrastructure or trade routes, to ensure that satellite images are used for risk analysis, and not for unauthorized surveillance.
- Apertus can help with developing scenarios and models that respect legal and ethical boundaries, using techniques like **data anonymization**, **synthetic data generation**, or **edge AI processing** to minimize sensitive data exposure.
- Create a platform where stakeholders (from grassroots to government) can input scenarios in a secure, collaborative environment based on trustable AI.
- The platform could include features like: secure processing of satellite data, with anonymization and compliance checks. AI-driven risk assessment based on the data could highlight potential supply chain disruptions.
- Allow stakeholders to **compare different scenarios**, identify patterns, and develop strategies for risk mitigation. Ensure that the models used are transparent and explainable to stakeholders, providing clear insights into the decision-making process.
- Engage with a diverse range of stakeholders (e.g., SMEs, NGOs, governmental agencies) to ensure the platform meets various needs.
- Collaborate on **data sharing protocols** that respect privacy while enabling collective risk analysis.
- Use the platform to share best practices in using satellite data for supply chain risk analysis, educating stakeholders on the legal and ethical use of satellite data.

---

### **Potential Next Steps**

- Develop a prototype of the scenario comparison platform using a limited set of satellite data and stakeholders.
- Test with a small group to validate the approach and identify areas for improvement.
- Organize a hackathon to bring together experts from different domains (finance, satellite data, AI, law, ethics) to refine the platform and address regulatory and ethical concerns.
- Explore funding opportunities (e.g., Swiss innovation grants, EU research programs) that support cross-disciplinary projects addressing supply chain resilience and AI for social good.
- Engage with experts in data privacy law, AI ethics, and satellite technology to ensure the solution respects all relevant regulations and ethical standards.

## Concluding thoughts

By leveraging Apertus and ALPS, this solution can help address the hurdle of surveillance capabilities and data protection regulations while enabling a collaborative platform for supply chain risk analysis. The platform would not only provide actionable insights but also set a precedent for responsible and transparent use of satellite data in financial risk analysis.

The use of satellite data in supply chain risk analysis can be both powerful and sensitive. By focusing on privacy-preserving AI, secure data handling, and stakeholder collaboration, we can create solutions that are both effective and responsible.

The Swiss AI ecosystem, through tools like Apertus and ALPS, provides the infrastructure and expertise to support such endeavors while respecting data protection and ethical considerations.

This approach can be a powerful example of how AI can be used to drive positive impact in global finance while respecting the complexities of data and regulation.

# :postal_horn:

Let me know if you'd like me to expand on any of these points or if you have further questions! --> oleg at datalets dot ch