/artificial intelligence

News and resources on artificial intelligence systems, innovations and initiatives worldwide.

AdviserSoftware and InterGen Data uncover key risks of DeepSeek

AdviserSoftware.com, part of FTRC, has collaborated with Robert Kirk, founder and CEO of InterGen Data, an AI-based financial lifetime events prediction service, to identify risks and recommendations for navigating DeepSeek.

  1 Be the first to comment

AdviserSoftware and InterGen Data uncover key risks of DeepSeek

Editorial

This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.

AI creates enormous opportunities for financial advisers, and interest in DeepSeek has been immense. To understand the potential risks better, AdviserSoftware and InterGen Data have analysed DeepSeek’s terms and conditions. They have identified the following risks and provide recommendations to advisers on how to navigate DeepSeek:

Key risks:

  • User liability for legal costs: Under Section 7.2 users must reimburse DeepSeek for legal fees, arbitration costs, attorney fees, travel expenses, and fines arising from their actions, including liabilities from third-party claims tied to alleged violations of laws, regulations, or platform rules.
  • Data privacy and government access: DeepSeek stores all user inputs (text, audio, files, chat history) and outputs on servers in China, subject to local cybersecurity laws requiring companies to share data with authorities upon request. IP addresses, device identifiers, keystroke patterns, and payment details are collected, raising privacy and compliance concerns, particularly regarding GDPR.
  • Content ownership and licensing: While users retain rights to their inputs, DeepSeek secures broad usage rights including non-exclusive, perpetual licenses to store, modify, and distribute user-generated content for commercial purposes. Outputs may be used to train future models, potentially exposing sensitive information shared in prompts.
  • Compliance and monitoring: Inputs/outputs are scanned for compliance, including risk filtering mechanisms where prohibited content ranges from hate speech to politically sensitive topics, with violations potentially leading to immediate suspension, removal of content, or bans without prior notice.
  • Third-party data sharing: SDKs are implemented from Chinese tech firms such as Tencent and ByteDance, potentially exposing user data to advertising partners, analytics platforms, and government-aligned entities under China’s data laws.

Recommendations:

  • Avoid sharing sensitive data: Refrain from inputting sensitive or confidential information by assuming all interactions with DeepSeek are monitored and stored indefinitely.
  • Consider jurisdictional risks: Be aware of data storage in China, especially in relation to strict data protection laws in the EU and UK.
  • Audit third-party integrations: SDKs such as Volcengine and Fengkongcloud increase exposure to Chinese infrastructure.
  • Understand licensing terms: Generated content may be repurposed commercially by DeepSeek, impacting confidentiality and proprietary information.

Speaking on this, Robert Kirk, founder and CEO of InterGen Data, commented: "Chasing the latest and greatest tech can be thrilling, but the hidden risks around data privacy, ownership, and liability exposure are significant and should not be overlooked. Firms must have a clear risk assessment framework and ensure their AI partners align with the highest standards of security, compliance, and ethical AI use.”

 "The T&Cs give DeepSeek the right to store information, which even extends as far as payment data. Advisers should understand the above issues in the context of any AI tool of service they are using. While the points identified above are not exhaustive, they identify valuable issues advisers should always understand and document before starting to use any technology service,” explained Ian McKenna, founder of AdviserSoftware.com.

Sponsored [Webinar] Conducting the payments orchestra: Why IT will drive future transaction banking models

Comments: (0)

[New Impact Study] Bank Legacy Transformation is Not a New Challenge: Exploring the SolutionsFinextra Promoted[New Impact Study] Bank Legacy Transformation is Not a New Challenge: Exploring the Solutions