DeepSeek: Difference between revisions
mNo edit summary |
→References: AT -> IA & Megalodon; Replace unarchivable Reuters with alternative and archive |
||
| (2 intermediate revisions by 2 users not shown) | |||
| Line 7: | Line 7: | ||
|Website=https://www.deepseek.com/ | |Website=https://www.deepseek.com/ | ||
|Logo=DeepSeek_logo.svg | |Logo=DeepSeek_logo.svg | ||
|Description=Chinese company developer of the DeepSeek large language models | |Description=Chinese company developer of the DeepSeek large language models that have privacy concerns for their data management collected from cloud services and mobile application. | ||
|CompanyAlias= Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. | |CompanyAlias= Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. | ||
}} | }} | ||
[[wikipedia:DeepSeek|'''DeepSeek''']] is a | [[wikipedia:DeepSeek|'''DeepSeek''']] is a Chinese artificial intelligence company founded on 17 July 2023 by Liang Wenfeng. It develops the [[wikipedia:DeepSeek_(chatbot)|DeepSeek large language model]] (LLM). | ||
==Consumer-impact summary== | ==Consumer-impact summary== | ||
===User Freedom=== | ===User Freedom=== | ||
In order to use the model from the official website or application, it is required to log in or create an account. However, | In order to use the model from the official website or application, it is required to log in or create an account. However, models can be downloaded and run locally. The code and the components can be modified as the models are distributed under the MIT license. | ||
The service provided from the official websites and applications offer unlimited conversations with the most recent models and features. | The service provided from the official websites and applications offer unlimited conversations with the most recent models and features. | ||
===User Privacy=== | ===User Privacy=== | ||
Using the chatbot from the official website, application or via the official API has raised several privacy concerns as how the company gathers and uses users data. Researches in South Korea found that the data collected is sent to [[ByteDance]]. <ref>https://web.archive.org/web/20250218135802/https://en.yna.co.kr/view/AEN20250218005300315</ref> | Using the chatbot from the official website, application or via the official API has raised several privacy concerns as how the company gathers and uses users data. Researches in South Korea found that the data collected is sent to [[ByteDance]].<ref>{{Cite web|url=https://en.yna.co.kr/view/AEN20250218005300315|title=DeepSeek sent S. Korean user data to China's ByteDance: regulator|first=Kim|last=Hyun-soo|date=2025-02-18|work=Yonhap News Agency|archive-url=https://web.archive.org/web/20250218135802/https://en.yna.co.kr/view/AEN20250218005300315|archivedate=2025-02-18|url-status=live}}</ref> | ||
===Business Model=== | ===Business Model=== | ||
DeepSeek grants access to the APIs of their models by fee | DeepSeek grants access to the APIs of their models by a fee for every 1M tokens. <ref>{{Cite web |publisher=DeepSeek|title=Models & Pricing |url=https://api-docs.deepseek.com/quick_start/pricing |url-status=live |archive-url=https://web.archive.org/web/20260308122355/https://api-docs.deepseek.com/quick_start/pricing |archive-date=2026-03-08 |access-date=13 Mar 2026 |website=DeepSeek Api Docs}}</ref> | ||
===Market Control=== | ===Market Control=== | ||
According to StatCounter as of February 2026, DeepSeek chatbot market share is 0.01% <ref>{{Cite web | According to StatCounter as of February 2026, DeepSeek chatbot market share is 0.01% <ref>{{Cite web |date=Feb 2026 |title=AI Chatbot Market Share Worldwide |url=https://gs.statcounter.com/ai-chatbot-market-share |url-status=live |archive-url=https://megalodon.jp/2026-0416-0436-09/https://gs.statcounter.com:443/ai-chatbot-market-share |archive-date=15 Apr 2026 |access-date=13 Mar 2026 |website=statcounter.com |publisher=StatCounter}}</ref>. | ||
==Censorship and biases== | ==Censorship and biases== | ||
[[File:DeepSeek_when_asked_about_Xi_Jinping_and_Narendra_Modi.png|thumb|A screenshot that shows a reply censorship when asked the model if Xi Jinping is an autocrat in nature. If it is asked the model the same question but with a non-chinese politician, it replies with a balanced answer]] | [[File:DeepSeek_when_asked_about_Xi_Jinping_and_Narendra_Modi.png|thumb|A screenshot that shows a reply censorship when asked the model if Xi Jinping is an autocrat in nature. If it is asked the model the same question but with a non-chinese politician, it replies with a more balanced answer]] | ||
The DeepSeek language models have heavy censorship of topics and information that is sensitive to the Chinese government, such as Tianmen Square Massacre, Taiwan's and Hong Kong's political statuses. If the user asks the model about information catalogued as "sensitive for the Chinese government", it will either block the reply and set a "Sorry, that's beyond my current scope. Let's talk about something else." pre-fabricated message or deny about it. {{Citation needed}} However, the censorship can be bypassed by adding additional instructions or by using the models from outside the chinese API. | The DeepSeek language models have heavy censorship of topics and information that is sensitive to the Chinese government, such as Tianmen Square Massacre, Taiwan's and Hong Kong's political statuses. If the user asks the model about information catalogued as "sensitive for the Chinese government", it will either block the reply and set a "Sorry, that's beyond my current scope. Let's talk about something else." pre-fabricated message or deny about it. {{Citation needed}} However, the censorship can be bypassed by adding additional instructions or by using the models from outside the chinese API. | ||
| Line 37: | Line 37: | ||
===Security breach (Jan. 2025)=== | ===Security breach (Jan. 2025)=== | ||
On January 29 2025, a group of cibersecurity | On January 29 2025, a group of cibersecurity researchers discovered over 1 million unencrypted sensitive records were leaked. This affected iOS users and information exposed include device information and conversation histories. | ||
<ref>https://www.forbes.com/sites/larsdaniel/2025/02/01/deepseek-data-leak-exposes--1000000-sensitive-records/</ref> | <ref>{{Cite web |last=Daniel |first=Lars |date=2025-02-01 |title=DeepSeek Data Leak Exposes 1 Million Sensitive Records |url=https://www.forbes.com/sites/larsdaniel/2025/02/01/deepseek-data-leak-exposes--1000000-sensitive-records/ |url-status=live |archive-url=https://web.archive.org/web/20260206122059/https://www.forbes.com/sites/larsdaniel/2025/02/01/deepseek-data-leak-exposes--1000000-sensitive-records/ |archive-date=6 Feb 2026 |work=Forbes}}</ref> | ||
<ref>https://www.quorumcyber.com/threat-intelligence/deepseek-app-exposes-sensitive-user-and-device-data-due-to-lack-of-encryption/</ref> | <ref>{{Cite web |date=2025-02-26 |title=DeepSeek App Exposes Sensitive User and Device Data Due to Lack of Encryption |url=https://www.quorumcyber.com/threat-intelligence/deepseek-app-exposes-sensitive-user-and-device-data-due-to-lack-of-encryption/ |url-status=live |archive-url=https://web.archive.org/web/20260317214033/https://www.quorumcyber.com/threat-intelligence/deepseek-app-exposes-sensitive-user-and-device-data-due-to-lack-of-encryption/ |archive-date=17 Mar 2026 |work=Quorum Cyber}}</ref> | ||
===Mobile application collecting user data (2025)=== | ===Mobile application collecting user data (2025)=== | ||
In 2025, | In 2025, Italy and Germany banned the DeepSeek mobile application from the [[Apple]] App Store and [[Google]] Play Store due to data management and protection not complying with the [[General Data Protection Regulation|GDPR]].<ref>{{Cite web |last1=Montti |first1=Roger |last2= |first2= |date=2025-06-27 |title=DeepSeek App Faces Ban In Germany For Illegal Transfer Of User Data |url=https://www.searchenginejournal.com/deepseek-app-faces-ban-in-germany-for-illegal-transfer-of-user-data/550172/ |url-status=live |archive-url=https://web.archive.org/web/20260213165658/https://www.searchenginejournal.com/deepseek-app-faces-ban-in-germany-for-illegal-transfer-of-user-data/550172/ |archive-date=13 Feb 2026 |work=Search Engine Journal}}</ref> <ref>{{Cite web |last1=Booth |first1=Robert |last2=Krupa |first2=Jakub |last3=Giuffrida |first3=Angela |date=2025-01-29 |title=DeepSeek blocked from some app stores in Italy amid questions on data use |url=https://www.theguardian.com/technology/2025/jan/29/deepseek-blocked-some-app-stores-italy-questions-data-use |url-status=live |archive-url=https://web.archive.org/web/20260215024119/https://www.theguardian.com/technology/2025/jan/29/deepseek-blocked-some-app-stores-italy-questions-data-use |archive-date=15 Feb 2026 |work=TheGuardian}}</ref> | ||
| Line 52: | Line 51: | ||
==See also== | ==See also== | ||
*[[ChatGPT]] | |||
*[[Claude]] | |||
*[[Google Gemini]] | |||
==References== | ==References== | ||
Latest revision as of 19:45, 15 April 2026
❗Article Status Notice: This Article is a stub
This article is underdeveloped, and needs additional work to meet the wiki's Content Guidelines and be in line with our Mission Statement for comprehensive coverage of consumer protection issues. Learn more ▼
DeepSeek
| Basic information | |
|---|---|
| Founded | 2023-07-17 |
| Legal Structure | Private |
| Industry | Artificial Intelligence |
| Also known as | Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. |
| Official website | https://www.deepseek.com/ |
DeepSeek is a Chinese artificial intelligence company founded on 17 July 2023 by Liang Wenfeng. It develops the DeepSeek large language model (LLM).
Consumer-impact summary
[edit | edit source]User Freedom
[edit | edit source]In order to use the model from the official website or application, it is required to log in or create an account. However, models can be downloaded and run locally. The code and the components can be modified as the models are distributed under the MIT license. The service provided from the official websites and applications offer unlimited conversations with the most recent models and features.
User Privacy
[edit | edit source]Using the chatbot from the official website, application or via the official API has raised several privacy concerns as how the company gathers and uses users data. Researches in South Korea found that the data collected is sent to ByteDance.[1]
Business Model
[edit | edit source]DeepSeek grants access to the APIs of their models by a fee for every 1M tokens. [2]
Market Control
[edit | edit source]According to StatCounter as of February 2026, DeepSeek chatbot market share is 0.01% [3].
Censorship and biases
[edit | edit source]
The DeepSeek language models have heavy censorship of topics and information that is sensitive to the Chinese government, such as Tianmen Square Massacre, Taiwan's and Hong Kong's political statuses. If the user asks the model about information catalogued as "sensitive for the Chinese government", it will either block the reply and set a "Sorry, that's beyond my current scope. Let's talk about something else." pre-fabricated message or deny about it. [citation needed] However, the censorship can be bypassed by adding additional instructions or by using the models from outside the chinese API.
An example of censorship occurs when the user sends a prompt containing "Taiwan is a country". This causes the model to claim that Taiwan is a "special administrative region of China" or to block the output if the model replies with a message confirming the region is a country. [citation needed]
Incidents
[edit | edit source]This is a list of all consumer-protection incidents this company is involved in. Any incidents not mentioned here can be found in the DeepSeek category.
Security breach (Jan. 2025)
[edit | edit source]On January 29 2025, a group of cibersecurity researchers discovered over 1 million unencrypted sensitive records were leaked. This affected iOS users and information exposed include device information and conversation histories. [4] [5]
Mobile application collecting user data (2025)
[edit | edit source]In 2025, Italy and Germany banned the DeepSeek mobile application from the Apple App Store and Google Play Store due to data management and protection not complying with the GDPR.[6] [7]
Products
[edit | edit source]- DeepSeek LLM
- DeepSeek Coder
- DeepSeek Math
See also
[edit | edit source]References
[edit | edit source]- ↑ Hyun-soo, Kim (2025-02-18). "DeepSeek sent S. Korean user data to China's ByteDance: regulator". Yonhap News Agency. Archived from the original on 2025-02-18.
- ↑ "Models & Pricing". DeepSeek Api Docs. DeepSeek. Archived from the original on 2026-03-08. Retrieved 13 Mar 2026.
- ↑ "AI Chatbot Market Share Worldwide". statcounter.com. StatCounter. Feb 2026. Archived from the original on 15 Apr 2026. Retrieved 13 Mar 2026.
- ↑ Daniel, Lars (2025-02-01). "DeepSeek Data Leak Exposes 1 Million Sensitive Records". Forbes. Archived from the original on 6 Feb 2026.
- ↑ "DeepSeek App Exposes Sensitive User and Device Data Due to Lack of Encryption". Quorum Cyber. 2025-02-26. Archived from the original on 17 Mar 2026.
- ↑ Montti, Roger (2025-06-27). "DeepSeek App Faces Ban In Germany For Illegal Transfer Of User Data". Search Engine Journal. Archived from the original on 13 Feb 2026.
- ↑ Booth, Robert; Krupa, Jakub; Giuffrida, Angela (2025-01-29). "DeepSeek blocked from some app stores in Italy amid questions on data use". TheGuardian. Archived from the original on 15 Feb 2026.