A Chinese-made artificial intelligence (AI) model called DeepSeek has shot to the top of Apple Store's downloads, spectacular investors and sinking some tech stocks.


Its latest version was released on 20 January, rapidly impressing AI professionals before it got the attention of the entire tech market - and the world.


US President Donald Trump stated it was a "wake-up call" for US companies who need to focus on "competing to win".


What makes DeepSeek so unique is the business's claim that it was built at a fraction of the cost of industry-leading models like OpenAI - due to the fact that it uses fewer advanced chips.


That possibility triggered chip-making huge Nvidia to shed practically $600bn (_ 482bn) of its market value on Monday - the most significant one-day loss in US history.


DeepSeek also raises questions about Washington's efforts to include Beijing's push for tech supremacy, offered that one of its key limitations has been a ban on the export of advanced chips to China.


Beijing, however, has doubled down, with President Xi Jinping declaring AI a top priority. And start-ups like DeepSeek are crucial as China rotates from standard production such as clothing and furniture to advanced tech - chips, electrical lorries and AI.


So what do we understand about DeepSeek?


Be cautious with DeepSeek, Australia states - so is it safe to utilize?


DeepSeek vs ChatGPT - how do they compare?


China's DeepSeek AI shakes industry and dents America's swagger


What is expert system?


AI can, sometimes, make a computer system look like a person.


A device utilizes the technology to find out and resolve problems, normally by being trained on massive quantities of info and identifying patterns.


The end result is software that can have conversations like an individual or anticipate people's shopping routines.


In recent years, it has become best referred to as the tech behind chatbots such as ChatGPT - and DeepSeek - likewise understood as generative AI.


These programs again learn from big swathes of data, consisting of online text and images, to be able to make brand-new material.


But these tools can produce frauds and often duplicate the predispositions included within their training information.


Countless people utilize tools such as ChatGPT to help them with everyday tasks like composing e-mails, summarising text, and addressing questions - and others even use them to assist with fundamental coding and studying.


DeepSeek is the name of a free AI-powered chatbot, which looks, feels and works quite like ChatGPT.


That indicates it's utilized for a lot of the very same tasks, though exactly how well it works compared to its competitors is up for argument.


It is apparently as powerful as OpenAI's o1 design - launched at the end of last year - in jobs consisting of mathematics and coding.


Like o1, R1 is a "reasoning" design. These designs produce responses incrementally, imitating a process comparable to how human beings reason through problems or concepts. It utilizes less memory than its rivals, ultimately decreasing the cost to perform jobs.


Like many other Chinese AI models - Baidu's Ernie or Doubao by ByteDance - DeepSeek is trained to avoid politically delicate concerns.


When the BBC asked the app what happened at Tiananmen Square on 4 June 1989, DeepSeek did not give any details about the massacre, a taboo subject in China.


It responded: "I am sorry, I can not answer that question. I am an AI assistant developed to provide valuable and safe responses."


Chinese federal government censorship is a huge obstacle for its AI goals internationally. But DeepSeek's base design appears to have actually been trained by means of precise sources while introducing a layer of censorship or withholding specific information through an extra safeguarding layer.


Deepseek states it has had the ability to do this inexpensively - researchers behind it claim it cost $6m (_ 4.8 m) to train, a portion of the "over $100m" alluded to by OpenAI employer Sam Altman when talking about GPT-4.


DeepSeek's creator apparently developed a shop of Nvidia A100 chips, which have been banned from export to China given that September 2022.


Some experts believe this collection - which some price quotes put at 50,000 - led him to build such a powerful AI model, by pairing these chips with more affordable, less advanced ones.


The same day DeepSeek's AI assistant became the most-downloaded totally free app on Apple's App Store in the US, it was struck with "massive destructive attacks", the company said, causing the business to short-lived limit registrations.


It was likewise hit by failures on its site on Monday.


Who lags DeepSeek?


DeepSeek was established in December 2023 by Liang Wenfeng, and launched its very first AI large language model the list below year.


Very little is understood about Liang, who finished from Zhejiang University with degrees in electronic information engineering and computer technology. But he now discovers himself in the international spotlight.


He was just recently seen at a meeting hosted by China's premier Li Qiang, showing DeepSeek's growing prominence in the AI market.


Unlike many American AI entrepreneurs who are from Silicon Valley, Mr Liang also has a background in financing.


He is the CEO of a hedge fund called High-Flyer, which utilizes AI to analyse financial information to make financial investment decisons - what is called quantitative trading. In 2019 High-Flyer became the first quant hedge fund in China to raise over 100 billion yuan ($13m).