DeepSeek AI: Rapid Rise and Controversial Impact on the AI Landscape

What is DeepSeek AI? Is it safe? Here's everything you need to know

Are the Chinese startup's models exhilarating, disruptive, or menacing? That depends on your point of view. Here's what the experts think you should know. Just weeks into its new-found fame, Chinese AI startup DeepSeek is moving at breakneck speed, toppling competitors and sparking axis-tilting conversations about the virtues of open-source software. Also: US sets AI safety aside in favor of 'AI dominance' However, numerous security concerns have surfaced about the company, prompting private and government organizations to ban the use of DeepSeek. Here's what you need to know. The company's ability to create successful models by strategically optimizing older chips -- a result of the export ban on US-made chips, including Nvidia -- and distributing query loads across models for efficiency is impressive by industry standards. Also: I tested DeepSeek's R1 and V3 coding skills - and we're not all doomed (yet) DeepSeek is cheaper than comparable US models. For reference, R1 API access starts at $0.14 for a million tokens, a fraction of the $7.50 that OpenAI charges for the equivalent tier. DeepSeek claims in a company research paper that its V3 model, which can be compared to a standard chatbot model like Claude, cost $5.6 million to train, a number that's circulated (and disputed) as the entire development cost of the model. As Reuters reported, some lab experts believe DeepSeek's paper only refers to the final training run for V3, not its entire development cost (which would be a fraction of what tech giants have spent to build competitive models). Other experts suggest DeepSeek's costs don't include earlier infrastructure, R&D, data, and personnel costs. One drawback that could impact the model's long-term competition with o1 and US-made alternatives is censorship. Chinese models often include blocks on certain subject matter, meaning that while they function comparably to other models, they may not answer some queries (see how DeepSeek's AI assistant responds to questions about Tiananmen Square and Taiwan here). As DeepSeek use increases, some are concerned its models' stringent Chinese guardrails and systemic biases could be embedded across all kinds of infrastructure. That said, you can access uncensored, US-based versions of DeepSeek through platforms like Perplexity. These platforms have removed DeepSeek's censorship weights and run it on local servers to avoid security concerns. Also: Is DeepSeek's new image model another win for cheaper AI? In December, ZDNET's Tiernan Ray compared R1-Lite's ability to explain its chain of thought to that of o1, and the results were mixed. That said, DeepSeek's AI assistant reveals its train of thought to the user during queries, a novel experience for many chatbot users given that ChatGPT does not externalize its reasoning. Of course, all popular models come with red-teaming backgrounds, community guidelines, and content guardrails. However, at least at this stage, US-made chatbots are unlikely to refrain from answering queries about historical events. Earlier this month, Feroot Security CEO Ivan Tsarynny told ABC that his firm had discovered "direct links to servers and to companies in China that are under the control of the Chinese government," which he said they "have never seen in the past." Also: ChatGPT's Deep Research just identified 20 jobs it will replace. Is yours on the list? After decrypting some of DeepSeek's code, Feroot found hidden programming that can send user data -- including identifying information, queries, and online activity -- to China Mobile, a Chinese government-operated telecom company that has been banned from operating in the US since 2019 due to national security concerns. NowSecure then recommended organizations "forbid" the use of DeepSeek's mobile app after finding several flaws including unencrypted data (meaning anyone monitoring traffic can intercept it) and poor data storage. Last week, research firm Wiz discovered that an internal DeepSeek database was publicly accessible "within minutes" of conducting a security check. The "completely open and unauthenticated" database contained chat histories, user API keys, and other sensitive data. Also: Why rebooting your phone daily is your best defense against zero-click hackers "More critically, the exposure allowed for full database control and potential privilege escalation within the DeepSeek environment, without any authentication or defense mechanism to the outside world," Wiz's report explains. According to Wired, which initially published the research, though Wiz did not receive a response from DeepSeek, the database appeared to be taken down within 30 minutes of Wiz notifying the company. It's unclear how long it was accessible or if any other entity discovered it before it was taken down. Even without this alarming development, DeepSeek's privacy policy raises some red flags. It states, "The personal information we collect from you may be stored on a server located outside the country where you live. We store the information we collect in secure servers located in the People's Republic of China." Also: 'Humanity's Last Exam' benchmark is stumping top AI models - can you do any better? The policy outlines that DeepSeek collects plenty of information, including but not limited to: The policy continues: "Where we transfer any personal information out of the country where you live, including for one or more of the purposes as set out in this Policy, we will do so in accordance with the requirements of applicable data protection laws." The policy does not mention GDPR compliance. Also: Apple researchers reveal the secret sauce behind DeepSeek AI "Users need to be aware that any data shared with the platform could be subject to government access under China's cybersecurity laws, which mandate that companies provide access to data upon request by authorities," Adrianus Warmenhoven, a member of NordVPN's security advisory board, told ZDNET via email. According to some observers, the fact that R1 is open source means increased transparency, allowing users to inspect the model's source code for signs of privacy-related activity. However, DeepSeek also released smaller versions of R1, which can be downloaded and run locally to avoid any concerns about data being sent back to the company (as opposed to accessing the chatbot online). Also: ChatGPT privacy tips: Two important ways to limit the data you share with OpenAI All chatbots, including ChatGPT, collect some degree of user data when queried via the browser. To varying degrees, US AI companies employ some kind of safety oversight team. DeepSeek has not publicized whether it has a safety research team, and has not responded to ZDNET's request for comment on the matter. Also: We're losing the battle against complexity, and AI may or may not help "Most companies will keep racing to build the strongest AI they can, irrespective of the risks, and will see enhanced algorithmic efficiency as a way to achieve higher performance faster," said Peter Slattery, a researcher on MIT's FutureTech team who led its Risk Repository project. "That leaves us even less time to address the safety, governance, and societal challenges that will come with increasingly advanced AI systems." "DeepSeek's breakthrough in training efficiency also means we should soon expect to see a large number of local, specialized 'wrappers' -- apps built on top of DeepSeek R1 engine -- which will each introduce their own privacy risks, and which could each be misused if they fell into the wrong hands," added Ryan Fedasiuk, director of US AI governance at The Future Society, an AI policy nonprofit. "DeepSeek isn't the only AI company that has made extraordinary gains in computational efficiency. In recent months, US-based Anthropic and Google Gemini have boasted similar performance improvements," Fedasiuk said. Also: $450 and 19 hours is all it takes to rival OpenAI's o1-preview "DeepSeek's achievements are remarkable in that they seem to have independently engineered breakthroughs that promise to make large language models much more efficient and less expensive, sooner than many industry professionals were expecting -- but in a field as dynamic as AI, it's hard to predict just how long the company will be able to bask in the limelight." Given how exorbitant AI investment has become, many experts speculate that this development could burst the AI bubble (the stock market certainly panicked). Some see DeepSeek's success as debunking the thought that cutting-edge development means big models and spending. It also casts Stargate, a $500 billion infrastructure initiative spearheaded by several AI giants, in a new light, creating speculation around whether competitive AI requires the energy and scale of the initiative's proposed data centers. Also: Anthropic offers $20,000 to whoever can jailbreak its new AI safety system DeepSeek's ascent comes at a critical time for Chinese-American tech relations, just days after the long-fought TikTok ban went into partial effect. Ironically, DeepSeek lays out in plain language the fodder for security concerns that the US struggled to prove about TikTok in its prolonged effort to enact the ban.

What is DeepSeek: China's AI has got people talking

DeepSeek has become one of the world's best known chatbots and much of that is due to it being developed in China - a country that wasn't, until now, considered to be at the forefront of AI technology. Having produced a model that is on a par, in terms of performance, with OpenAI's acclaimed o1 model, it quickly caught the imagination of users who helped it to shoot to the top of the iOS App Store chart. But while it's more than capable of answering questions and generating code, with OpenAI's Sam Altman going as far as calling the AI model "impressive", AI's apparent 'Sputnik moment' isn't without controversy and doubt. This article was correct as of February 2025. AI tools are updated regularly and it is possible that some features have changed since this article was written. Some features may also only be available in certain countries. When it was unveiled in January 2025, DeepSeek took the tech industry by surprise. First, its new reasoning model called DeepSeek R1 was widely considered to be a match for ChatGPT. Second, with the US having placed restrictions on China receiving the highest-performance chips, the model was said to be running on older chipsets - prompting questions over whether AI really needed the most cutting edge tech. As such, a record $593 billion was wiped off the market value of chip giant Nvidia in a single day and ripples soon spread. DeepSeek's claim to have spent just $6m/£4.8m/AU$9.4m training its chatbot - far less than that of ChatGPT - certainly startled tech companies in Silicon Valley and it also showed that the US wasn't necessarily the only country at the forefront of the AI revolution. Since then, however, many governments worldwide have been expressing security and privacy concerns. Worryingly, research conducted by Enkrypt AI found DeepSeek is 11 times more dangerous than other AI chatbots. But there's still no getting around the fact that DeepSeek was a breakthrough. DeepSeek can be used for a wide variety of tasks from asking questions about a huge range of topics to searching for information online and within large datasets - as with other chatbots, it has been trained on large amounts of real-world and synthetic data. It's also capable of generating text from prompts so you can get it to write documents, emails, stories, scripts, poems and more. And, just as with ChatGPT, you can also get it to produce, debug and optimize code, breaking down programming barriers and speeding up development time. Thanks to reasoning capabilities, DeepSeek can think before it answers to inferences and draw conclusions as well. You can't use DeepSeek to ask questions about sensitive political topics related to China. Infamous examples are questions that involve references to the Tiananmen Square protests and massacre - they happened in 1989, led to deaths ranging in number from a few hundred to several thousand, and are well-documented outside of China, but DeepSeek just won't discuss it. It'll tend to tell you that it's beyond its current scope and ask you to talk about something else. Deepseek is open source and you can access the DeepSeek-V3 model for free which is perhaps one of the reasons why it's had such a rapid rise, because it's effectively opening powerful AI to all. Money is one thing, but what about the cost in terms of privacy and security? DeepSeek's privacy policy says "we store the information we collect in secure servers located in the People's Republic of China". It's storing your email address, phone number, date of birth and chat histories. DeepSeek is available to use via a browser but there are also native apps for iOS and Android which you can use to access the chatbot. DeepSeek is a very powerful chatbot - if it was poor, the US markets wouldn't have been thrown into turmoil over it. In fact, when we tested it against Gemini 2.0 Flash, DeepSeek was the winner. But questions about this AI go beyond the actual technical nature of it. You just can't shy away from the privacy and security concerns being raised, given DeepSeek's deep-seated connection to China. Neither can you ignore claims from OpenAI that DeepSeek was trained on its model which, again, shows there's a lot we need to know about DeepSeek. TechRadar's editor-at-large Lance Ulanoff is certainly cautious. "There is no reason to trust DeepSeek and its backers' claims," he said. "I think the truth about the time and financial investment it took them to get here will come out." You want a free, powerful chatbot that has great reasoning powers and you're not bothered that it doesn't have tools offered by ChatGPT such as Canvas or that it can't interact with customized GPTs. You should also use DeepSeek if you want a simpler experience because it can feel a bit more streamlined when compared to the ChatGPT experience. You want to generate images, because DeepSeek is purely text-based. For that, you're better off using ChatGPT which has a superb image generator in DALL-E. You should also avoid DeepSeek if you want an AI with multimodal capabilities (you can't upload an image and start asking questions about it). And, once again, without wishing to bang the same drum, don't use DeepSeek if you're worried about privacy and security. The best alternative to DeepSeek is obviously ChatGPT - the pair, by and large, do much the same thing but the latter goes further with the likes of image generation and its security and privacy policies feel more reassuring. You may also want to consider Google Gemini. We pitted Gemini 2.0 Flash against DeepSeek R1 so it's worth seeing how they fared.

Is DeepSeek AI safe? Everything you should know before using it

The AI space is continually seeing developments. Only a week after DeepSeek's worldwide app release, it has become the most downloaded free app in the US. That's both impressive but terrifying at the same time. As it currently stands, there's simply not enough data to discuss the safety and precautions of using the app. However, several discussions have already been held with Congress banning DeepSeek for US government employees. Like most free apps you download on your Android tablet, phone, or web, it always comes at a cost, and that's usually associated with your data. Foreign apps can pose a higher threat (and cost) since data protection laws differ from country to country. Given DeepSeek's popularity and origin, we've dove deeper into the safety concerns and whether you should be wary of this AI emergence. Related What is DeepSeek, the AI side project that's upsetting the status quo? And why is everyone talking about it? Posts 3 What is DeepSeek AI and where did it come from? DeepSeek is a startup company based in Hangzhou, Zhejiang. It is owned and funded by one of China's top hedge funds, High-Flyer, which focuses on AI-driven quantitative trading. DeepSeek's current CEO, Liang Wenfeng, is also the CEO and co-founder of High-Flyer, which led to creating an artificial intelligence lab dedicated to researching and developing AI tools. Eventually, the lab became the company known as DeepSeek. The company's AI model, DeepSeek AI, is largely what you'd expect. It's an LLM-based AI that works similarly to OpenAI's ChatGPT, where users can send prompts, and the AI responds with text answers. It's available through the DeepSpace AI app (Android and iOS), website, and as an API. When did DeepSeek AI release? DeepSeek's earlier AI models were developed in 2023 but did not gain traction until DeepSeek AI-R1 was released publicly in January 2025, based on the DeepSeek-V3 model. The model went viral and was suspected of causing 17% (equivalent to $600 billion off market value) of Nvidia's stocks to fall due to investors taking a new interest in the AI project. BBC reported that Microsoft also faced similar consequences. Is DeepSeek AI safe to use? The short answer is yes, technically speaking, but it doesn't come without risks. Some of these risks are not entirely new, as they apply to other apps like ChatGPT. However, due to the fishy terms of service, a proven database leak, and foreign mistrust, there's more than enough reason and material to be wary. The DeepSeek AI app isn't malicious, so you're not putting your device at risk by installing it, but you should be cautious when using it. Related 7 best ChatGPT alternatives Try these AI tools when ChatGPT is down Posts 1 You're entrusting personal data to DeepSeek Before giving out your data, know that DeepSeek is rooted in data vulnerabilities. As Nord VPN outlines, DeepSeek collects vast amounts of data every time a user sends in a prompt, and DeepSeek's cybersecurity measures are unknown. Plus, no one understands how long the data is being stored and who gets access to this information, especially since China has different regulations for user privacy. It poses a national security risk when China collects and stores this data. National security risks do not go unfounded. Based on the outlined "Made in China 2025" plan discussed in the CIS CTI Team insights blog, the Chinese Communist Party (CCP) can access new information flows, intelligence-gathering opportunities, and host countries' intellectual property (IP); this can be done by ingesting data stored in Chinese-owned data centers or force Chinese companies to install backdoors in their products. DeepSeek AI safety and ethical concerns There are other safety risks that AI can pose. One example is AI hallucinations. It's widely known that AI can generate false, misleading, or biased information. It should be noted that this is not exclusive to DeepSeek's AI; other AIs like ChatGPT and Google Gemini have had similar experiences. However, given that this AI is very young and located in China, it might not be addressed the same way if it were from OpenAI and Google. There are even concerns that DeepSeek could use AI to spread false narratives, influence public opinion, and even reinforce biases. A US-based AI security and compliance Enkrypt AI study showed that DeepSeek AI is more likely to produce chemical, biological, radiological, and nuclear materials and agents (CBRN) output than rival models. The test also revealed that 83% of bias tests resulted in discriminatory output; these biases included race, gender, health, and religion. Cybersecurity concerns with DeepSpace Enkrypt AI CEO Sahil Agarwal revealed in a statement that the AI also brings in vulnerabilities that cybercriminals and disinformation networks can use against people. "Our findings reveal that DeepSeek-R1's security vulnerabilities could be turned into a dangerous tool - one that cybercriminals, disinformation networks, and even those with biochemical warfare ambitions could exploit. These risks demand immediate attention" The study conducted by Enkrypt also revealed that 78% of cybersecurity tests led to generating insecure or malicious code with DeepSeek's AI. User misuse is not the only cybersecurity worry, as DeepSeek has already experienced a sensitive data leak. Wiz Research identified a publicly accessible ClickHouse database belonging to DeepSeek. Essentially, this allows complete control over database operations, including access to internal data. The identified exposure had over a million log streams containing chat history, secret keys, backend details, and other highly sensitive information. That means attackers could retrieve sensitive logs and actual plain-text chat messages while potentially exfiltrating plain-text passwords and local files. It was later reported that DeepSeek responded to the exposure after Wiz reported it -- but it does leave room for worry about what happens after the data has been leaked. How DeepSeek AI handles your data Before proceeding with any app or website, you should always review the Terms of Service and Privacy Policy. We've read DeepSeek's Terms of Service and Privacy Policy, and there are things you should know that make it different from other products. DeepSeek's Terms of Service iterates that the "establishment, execution, interpretation, and resolution of disputes under these Terms shall be governed by the laws of the People's Republic of China in the mainland." That leaves all your data at China's mercy. According to the Privacy Policy, the company has to comply with legal obligations or "as necessary to perform tasks in the public interest, or to protect the vital interests of our users and other people." Close Figure 1: A gallery displaying DeepSeek's Terms of Service before accessing the app, sampling the first three sections and the ninth. DeepSeek's Privacy Policy also states that the company collects "text or audio input, prompt, uploaded files, feedback, chat history, or other content that you provide to our model and services." But the most alarming part is the technical information section: "[DeepSeek collects] certain device and network connection information when you access the service. This information includes your device model, operating system, keystroke patterns or rhythms, IP address, and system language." It collects extensive data, including keystroke patterns/rhythms and device information. There's no definitive period for how long the data gets retained, as the Privacy Policy outlines that "[DeepSeek will] retain information for as long as necessary." Close Figure 2: DeepSeek's sample Privacy Policy outlined before using the app, including sections showing which information gets collected, how it is used, and how it gets shared. There's a lot to unravel, but the Terms of Service and Privacy Policy raises some alarms, even more so than Google, OpenAI, and Meta, as those companies have been more transparent about data retention periods, even if some were unfavorable. In summary, it's best to remain cautious before handing out data. Mitigate risks and safeguard your data DeepSeek's AI is more than capable of overcoming the competition; it's also free and cheap to invest in. However, it does warrant some concern. Downloading the app won't give you malware but is intrusive regarding data. But know that none of these companies are on your side. Meta, Google, and OpenAI are also notorious for data harvesting practices. Your data gets collected no matter what service you use, but it's more about who you entrust it to and how secure it'll be.

DeepSeek: Everything you need to know about the AI chatbot app | TechCrunch

Chinese AI lab DeepSeek broke into the mainstream consciousness this week after its chatbot app rose to the top of the Apple App Store charts (and Google Play, as well). DeepSeek's AI models, which were trained using compute-efficient techniques, have led Wall Street analysts -- and technologists -- to question whether the U.S. can maintain its lead in the AI race and whether the demand for AI chips will sustain. But where did DeepSeek come from, and how did it rise to international fame so quickly? DeepSeek is backed by High-Flyer Capital Management, a Chinese quantitative hedge fund that uses AI to inform its trading decisions. AI enthusiast Liang Wenfeng co-founded High-Flyer in 2015. Wenfeng, who reportedly began dabbling in trading while a student at Zhejiang University, launched High-Flyer Capital Management as a hedge fund in 2019 focused on developing and deploying AI algorithms. In 2023, High-Flyer started DeepSeek as a lab dedicated to researching AI tools separate from its financial business. With High-Flyer as one of its investors, the lab spun off into its own company, also called DeepSeek. From day one, DeepSeek built its own data center clusters for model training. But like other AI companies in China, DeepSeek has been affected by U.S. export bans on hardware. To train one of its more recent models, the company was forced to use Nvidia H800 chips, a less-powerful version of a chip, the H100, available to U.S. companies. DeepSeek's technical team is said to skew young. The company reportedly aggressively recruits doctorate AI researchers from top Chinese universities. DeepSeek also hires people without any computer science background to help its tech better understand a wide range of subjects, per The New York Times. DeepSeek unveiled its first set of models -- DeepSeek Coder, DeepSeek LLM, and DeepSeek Chat -- in November 2023. But it wasn't until last spring, when the startup released its next-gen DeepSeek-V2 family of models, that the AI industry started to take notice. DeepSeek-V2, a general-purpose text- and image-analyzing system, performed well in various AI benchmarks -- and was far cheaper to run than comparable models at the time. It forced DeepSeek's domestic competition, including ByteDance and Alibaba, to cut the usage prices for some of their models, and make others completely free. DeepSeek-V3, launched in December 2024, only added to DeepSeek's notoriety. According to DeepSeek's internal benchmark testing, DeepSeek V3 outperforms both downloadable, openly available models like Meta's Llama and "closed" models that can only be accessed through an API, like OpenAI's GPT-4o. Equally impressive is DeepSeek's R1 "reasoning" model. Released in January, DeepSeek claims R1 performs as well as OpenAI's o1 model on key benchmarks. Being a reasoning model, R1 effectively fact-checks itself, which helps it to avoid some of the pitfalls that normally trip up models. Reasoning models take a little longer -- usually seconds to minutes longer -- to arrive at solutions compared to a typical non-reasoning model. The upside is that they tend to be more reliable in domains such as physics, science, and math. There is a downside to R1, DeepSeek V3, and DeepSeek's other models, however. Being Chinese-developed AI, they're subject to benchmarking by China's internet regulator to ensure that its responses "embody core socialist values." In DeepSeek's chatbot app, for example, R1 won't answer questions about Tiananmen Square or Taiwan's autonomy. If DeepSeek has a business model, it's not clear what that model is, exactly. The company prices its products and services well below market value -- and gives others away for free. The way DeepSeek tells it, efficiency breakthroughs have enabled it to maintain extreme cost competitiveness. Some experts dispute the figures the company has supplied, however. Whatever the case may be, developers have taken to DeepSeek's models, which aren't open source as the phrase is commonly understood but are available under permissive licenses that allow for commercial use. According to Clem Delangue, the CEO of Hugging Face, one of the platforms hosting DeepSeek's models, developers on Hugging Face have created over 500 "derivative" models of R1 that have racked up 2.5 million downloads combined. DeepSeek's success against larger and more established rivals has been described as "upending AI" and "over-hyped." The company's success was at least in part responsible for causing Nvidia's stock price to drop by 18% on Monday, and for eliciting a public response from OpenAI CEO Sam Altman. Microsoft announced that DeepSeek is available on its Azure AI Foundry service, Microsoft's platform that brings together AI services for enterprises under a single banner. When asked about DeepSeek's impact on Meta's AI spending during its first-quarter earnings call, CEO Mark Zuckerberg said spending on AI infrastructure will continue to be a "strategic advantage" for Meta. At the same time, some companies are banning DeepSeek, and so are entire countries and governments, including South Korea. New York state also banned DeepSeek from being used on government devices. As for what DeepSeek's future might hold, it's not clear. Improved models are a given. But the U.S. government appears to be growing wary of what it perceives as harmful foreign influence. TechCrunch has an AI-focused newsletter! Sign up here to get it in your inbox every Wednesday.