Episode Summary: In this episode, we delve into the rapidly growing popularity of DeepSeek, a Chinese AI model, and uncover the potential security and privacy risks it poses. From data transmission vulnerabilities to troubling terms of service, we explore the reasons why caution is key when considering this innovative AI tool. Episode Notes: • Introduction: ◦ DeepSeek's rapid rise to prominence and its potential impact on the AI landscape1.... ◦ Initial praise for its efficiency and advanced AI capabilities2. • Data Privacy Concerns: ◦ DeepSeek's data storage in China raises concerns about state surveillance and compliance with global privacy standards5.... ◦ Collection of extensive user data, including device details, IP addresses, and usage patterns5.... • Terms of Service Red Flags: ◦ Troubling aspects of DeepSeek's ToS, allowing data collection on usage, prompting, device, network, and personal activity7. ◦ Comparison to data collection practices of companies like Google, Apple, and Microsoft, but with the added concern of data handling in China7. • Security Vulnerabilities: ◦ Unencrypted data transmission: The DeepSeek iOS app transmits sensitive user and device data without encryption, exposing it to interception and manipulation attacks8.... ◦ Weak encryption practices: Use of outdated encryption algorithms like 3DES with hard-coded keys, making it vulnerable to cryptographic attacks5.... ◦ Database exposure: Publicly accessible database linked to DeepSeek exposed chat histories, API secrets, and backend operational details14.... ◦ Cyberattack target: DeepSeek has already suffered significant cyberattacks, making it an attractive target for malicious actors15.... • Global Response: ◦ Bans and warnings issued by various countries and organizations, including the U.S. Navy, NASA, and government agencies in Italy and Taiwan18.... ◦ U.S. lawmakers' efforts to restrict DeepSeek's use on government devices3.... • Mitigation Strategies: ◦ Guidance to avoid inputting sensitive information into any LLMs that aren't self-hosted23.... ◦ Running open-source models locally to reduce risks, while being mindful of potential vulnerabilities25.... ◦ Using network monitoring tools like Wireshark to observe data transmission26.... ◦ Implementing robust mobile security solutions like Approov to ensure only legitimate app instances communicate with backend services32.... • Expert Perspectives: ◦ Discussion of DeepSeek's potential for misuse, including the generation of ransomware development scripts34.... ◦ Analysis of the balance between security, privacy, and the benefits of AI innovation36. • Practical Advice: ◦ Researching and understanding the terms of service before using any new app7.... ◦ Being cautious about free apps, recognizing that "if it's free, you're the product"36. ◦ Staying informed about the latest cybersecurity risks and data privacy concerns36. Sponsor Message: This episode is brought to you by Approov (<a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.approov.io" target="_blank">https://www.approov.io</a>). In today's mobile-first world, securing your apps is more critical than ever. Approov's mobile app attestation ensures that only genuine and unmodified versions of your app can access your backend services, protecting against fraud, data breaches, and unauthorized access. Learn more about how Approov can safeguard your mobile applications at <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.approov.io" target="_blank">approov.io</a>. Approov is the only commercially available mobile app attestation solution that operates across iOS, Android (both GMS and non-GMS), and HarmonyOS platforms33. Keywords: DeepSeek, AI, cybersecurity, data privacy, China, mobile security, app security, threat actors, data breaches, Approov, app attestation, LLM, infostealer, surveillance, terms of service, data exfiltration, open source, mobile app, VPN Links to Source Materials: • Are there any legitimate security concerns regarding DeepSeek?: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.reddit.com%2Fr%2Fcybersecurity%2Fcomments%2F1icxzb3%2Fare_there_any_legitimate_security_concerns%2F" target="_blank">https://www.reddit.com/r/cybersecurity/comments/1icxzb3/are_there_any_legitimate_security_concerns/</a> • DeepSeek App Transmits Sensitive User and Device Data Without Encryption: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fthehackernews.com%2F2025%2F02%2Fdeepseek-app-transmits-sensitive-user.html" target="_blank">https://thehackernews.com/2025/02/deepseek-app-transmits-sensitive-user.html</a> • Experts Flag Security, Privacy Risks in DeepSeek AI App: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fkrebsonsecurity.com%2F2025%2F02%2Fexperts-flag-security-privacy-risks-in-deepseek-ai-app%2F" target="_blank">https://krebsonsecurity.com/2025/02/experts-flag-security-privacy-risks-in-deepseek-ai-app/</a> • Wiz Research Uncovers Exposed DeepSeek Database Leaking: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.wiz.io%2Fblog%2Fwiz-research-uncovers-exposed-deepseek-database-leak" target="_blank">https://www.wiz.io/blog/wiz-research-uncovers-exposed-deepseek-database-leak</a> • The Hidden Risks of DeepSeek AI: Why Caution Is Key | SBS: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fsbscyber.com%2Fblog%2Fdeepseek-ai-dangers" target="_blank">https://sbscyber.com/blog/deepseek-ai-dangers</a> • The Independent: Feroot Security Uncovers DeepSeek’s Hidden Code Sending User Data to China - Feroot Security: <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.feroot.com%2Fnews%2Fthe-independent-feroot-security-uncovers-deepseeks-hidden-code-sending-user-data-to-china%2F" target="_blank">https://www.feroot.com/news/the-independent-feroot-security-uncovers-deepseeks-hidden-code-sending-user-data-to-china/</a> Call to Action: What are your thoughts on DeepSeek? Share your concerns and experiences with us on social media using #UpwardlyMobile #DeepSeek. Credits: • Host: [Your Name] • Editor: [Editor's Name] • Sponsor: Approov (<a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.approov.io" target="_blank">approov.io</a>)

Episode Summary: In this episode, we delve into the rapidly growing popularity of DeepSeek, a Chinese AI model, and uncover the potential security and privacy risks it poses. From data transmission vulnerabilities to troubling terms of service, we...

Episode Summary: In this episode, we delve into the rapidly growing popularity of DeepSeek, a Chinese AI model, and uncover the potential security and privacy risks it poses. From data transmission vulnerabilities to troubling terms of service, we explore the reasons why caution is key when considering this innovative AI tool. Episode Notes: • Introduction: ◦ DeepSeek's rapid rise to prominence and its potential impact on the AI landscape1.... ◦ Initial praise for its efficiency and advanced AI capabilities2. • Data Privacy Concerns: ◦ DeepSeek's data storage in China raises concerns about state surveillance and compliance with global privacy standards5.... ◦ Collection of extensive user data, including device details, IP addresses, and usage patterns5.... • Terms of Service Red Flags: ◦ Troubling aspects of DeepSeek's ToS, allowing data collection on usage, prompting, device, network, and personal activity7. ◦ Comparison to data collection practices of companies like Google, Apple, and Microsoft, but with the added concern of data handling in China7. • Security Vulnerabilities: ◦ Unencrypted data transmission: The DeepSeek iOS app transmits sensitive user and device data without encryption, exposing it to interception and manipulation attacks8.... ◦ Weak encryption practices: Use of outdated encryption algorithms like 3DES with hard-coded keys, making it vulnerable to cryptographic attacks5.... ◦ Database exposure: Publicly accessible database linked to DeepSeek exposed chat histories, API secrets, and backend operational details14.... ◦ Cyberattack target: DeepSeek has already suffered significant cyberattacks, making it an attractive target for malicious actors15.... • Global Response: ◦ Bans and warnings issued by various countries and organizations, including the U.S. Navy, NASA, and government agencies in Italy and Taiwan18.... ◦ U.S. lawmakers' efforts to restrict DeepSeek's use on government devices3.... • Mitigation Strategies: ◦ Guidance to avoid inputting sensitive information into any LLMs that aren't self-hosted23.... ◦ Running open-source models locally to reduce risks, while being mindful of potential vulnerabilities25.... ◦ Using network monitoring tools like Wireshark to observe data transmission26.... ◦ Implementing robust mobile security solutions like Approov to ensure only legitimate app instances communicate with backend services32.... • Expert Perspectives: ◦ Discussion of DeepSeek's potential for misuse, including the generation of ransomware development scripts34.... ◦ Analysis of the balance between security, privacy, and the benefits of AI innovation36. • Practical Advice: ◦ Researching and understanding the terms of service before using any new app7.... ◦ Being cautious about free apps, recognizing that "if it's free, you're the product"36. ◦ Staying informed about the latest cybersecurity risks and data privacy concerns36. Sponsor Message: This episode is brought to you by Approov (<a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.approov.io" target="_blank">https://www.approov.io</a>). In today's mobile-first world, securing your apps is more critical than ever. Approov's mobile app attestation ensures that only genuine and unmodified versions of your app can access your backend services, protecting against fraud, data breaches, and unauthorized access. Learn more about how Approov can safeguard your mobile applications at <a href="https://www.google.com/url?sa=E&amp;q=https%3A%2F%2Fwww.approov.io" target="_blank">approov.io</a>. Approov is the only commercially available mobile app attestation solution that operates...

Is DeepSeek Safe? Navigating the Risks of the Chinese AI Model