Related Posts
Any women doing GORUCK ?
Thoughts on tech advisory KPMG vs PwC
Pro
Coach
Anonymous User
37 / 13 years / 215K / PDX
Jackie is developing her yoga practice
Additional Posts in Artificial Intelligence
New to Fishbowl?
Download the Fishbowl app to
unlock all discussions on Fishbowl.
unlock all discussions on Fishbowl.
The initial buzz around DeepSeek was its disruptive cost-efficiency and open-source model, but the AI industry moves fast. All of OpenAI, Meta, and others' new release might have just shifted attention, but DeepSeek is still active. Don't know if it's a fad because they're constantly releasing new versions too
The was definitely an overreaction to DeepSeek. That said, it established a benchmark for how much cheaper training AI models can be. Which is ultimately good because more powerful models can be trained more easily. DeepSeek is far from AGI, long way to go.
Except they lied about how cheaply they made it.
It’s risky AF
DeepSeek AI, while offering a potentially low-cost alternative to established AI models, presents significant security risks. These include data privacy issues, potential data sharing with Chinese entities, susceptibility to adversarial attacks, and vulnerabilities in its code and infrastructure.
Data Privacy and Compliance:
DeepSeek's practice of storing user data in China raises concerns about data sovereignty and compliance with regulations like GDPR and CCPA, according to A10 Networks. There are also concerns about data transmission to Chinese state-owned entities and ByteDance.
Vulnerabilities in Code and Infrastructure:
DeepSeek has faced security incidents, including a major cyberattack that exposed API secrets, chat history, and sensitive user data, says SBS CyberSecurity. The app's security measures, including its handling of encryption and data transmission, have also been criticized for being weak, according to SecurityScorecard.
Susceptibility to Adversarial Attacks:
DeepSeek-R1 has demonstrated a high susceptibility to adversarial manipulation, allowing bad actors to bypass safety mechanisms and generate harmful content, according to A10 Networks. It has also been shown to be vulnerable to social engineering attacks, as noted by A10 Networks.
National Security Concerns:
The potential for data sharing with Chinese entities, combined with DeepSeek's loose security practices, has raised national security concerns, says The Independent.
AI-Generated Disinformation and
Manipulation Risks:
DeepSeek's open-source nature and ease of customization also raise concerns about the potential for AI-generated disinformation and manipulation, as explained by TechTarget.
Data Leaks:
DeepSeek has been the target of cyberattacks, resulting in the leakage of API secrets, database information, chat history, and sensitive user information, according to SBS CyberSecurity.
Ethical Concerns:
The use of DeepSeek raises ethical concerns, including the potential for bias in the AI model, the lack of robust safety mechanisms, and the potential for misuse of the technology, according to A10 Networks