Latest news with #DEEPSEEK
Mint
8 hours ago
- Business
- Mint
There is a vast hidden workforce behind AI
WHEN DEEPSEEK, a hotshot Chinese firm, released its cheap large language model late last year it overturned long-standing assumptions about what it will take to build the next generation of artificial intelligence (AI). This will matter to whoever comes out on top in the epic global battle for AI supremacy. Developers are now reconsidering how much hardware, energy and data are needed. Yet another, less discussed, input in machine intelligence is in flux too: the workforce. To the layman, AI is all robots, machines and models. It is a technology that kills jobs. In fact, there are millions of workers involved in producing AI models. Much of their work has involved tasks like tagging objects in images of roads in order to train self-driving cars and labelling words in the audio recordings used to train speech-recognition systems. Technically, annotators give data the contextual information computers need to work out the statistical associations between components of a dataset and their meaning to human beings. In fact, anyone who has completed a CAPTCHA test, selecting photos containing zebra crossings, may have inadvertently helped train an AI. This is the 'unsexy" part of the industry, as Alex Wang, the boss of Scale AI, a data firm, puts it. Although Scale AI says most of its contributor work happens in America and Europe, across the industry much of the labour is outsourced to poor parts of the world, where lots of educated people are looking for work. The Chinese government has teamed up with tech companies, such as Alibaba and to bring annotation jobs to far-flung parts of the country. In India the IT industry body, Nasscom, reckons annotation revenues could reach $7bn a year and employ 1m people there by 2030. That is significant, since India's entire IT industry is worth $254bn a year (including hardware) and employs 5.5m people. Annotators have long been compared to parents, teaching models and helping them make sense of the world. But the latest models don't need their guidance in the same way. As the technology grows up, are its teachers becoming redundant? Data annotation is not new. Fei Fei Li, an American computer scientist known as 'the godmother of AI", is credited with firing the industry's starting gun in the mid-2000s when she created ImageNet, the largest image dataset at the time. Ms Li realised that if she paid college students to categorise the images, which was then how most researchers did things, the task would take 90 years. Instead, she hired workers around the world using Mechanical Turk, an online gig-work platform run by Amazon. She got some 3.2m images organised into a dataset in two and a half years. Soon other AI labs were outsourcing annotation work this way, too. Over time developers got fed up with the low-quality annotation done by untrained workers on gig-work sites. AI-data firms, such as Sama and iMerit, emerged. They hired workers across the poor world. Informal annotation work continued but specialist platforms emerged for AI work, like those run by Scale AI, which tests and trains workers. The World Bank reckons that between 4.4% and 12.4% of the global workforce is involved in gig work, including annotation for AI. Krystal Kauffman, a Michigan resident who has been doing data work online for a decade, reckons that tech companies have an interest in keeping this workforce hidden. 'They are selling magic—this idea that all these things happen by themselves," Ms Kauffman, says. 'Without the magic part of it, AI is just another product." A debate in the industry has been about the treatment of the workers behind AI. Firms are reluctant to share information on wages. But American annotators generally consider $10-20 per hour to be decent pay on online platforms. Those in poor countries often get $4-8 per hour. Many must use monitoring tools that track their computer activity and are penalised for being slow. Scale AI has been hit with several lawsuits over its employment practices. The firm denies wrongdoing and says: 'We plan to defend ourselves vigorously." The bigger issue, though, is that basic annotation work is drying up. In part, this was inevitable. If AI was once a toddler who needed a parent to point things out and to help it make sense of the world around it, the technology has grown into an adolescent who needs occasional specialist guidance and advice. AI labs increasingly use pre-labelled data from other AI labs, which use algorithms to apply labels to datasets. Take the example of self-driving tractors developed by Blue River Technology, a subsidiary of John Deere, an agricultural-equipment giant. Three years ago the group's engineers in America would upload pictures of farmland into the cloud and provide iMerit staff in Hubli, India, with careful instructions on what to label: tractors, buildings, irrigation equipment. Now the developers use pre-labelled data. They still need iMerit staff to check that labelling and to deal with 'edge cases", for example where a dust cloud obscures part of the landscape or a tree throws shade over crops, confusing the model. A process that took months now takes weeks. From baby steps The most recent wave of AI models has changed data work more dramatically. Since 2022, when OpenAI first let the public play with its ChatGPT chatbot, there has been a rush of interest in large language models. Data from Pitchbook, a research firm, suggest that global venture-capital funding for AI startups jumped by more than 50% in 2024 to $131.5bn, even as funding for other startups fell. Much of it is going into newer techniques for developing AI, which do not need data annotated in the same way. Iva Gumnishka at Humans in the Loop, a social enterprise, says firms doing low-skilled annotation for older computer-vision and natural-language-processing clients are being 'left behind". There is still demand for annotators, but their work has changed. As businesses start to deploy AI, they are building smaller specialised models and looking for highly educated annotators to help. It has become fairly common for adverts for annotation jobs to require a PhD or skills in coding and science. Now that researchers are trying to make AI more multilingual, demand for annotators who speak languages other than English is growing, too. Sushovan Das, a dentist working on medical-AI projects at iMerit, reckons that annotation work will never disappear. 'This world is constantly evolving," he says. 'So the AI needs to be improved time and again." New roles for humans in training AI are emerging. Epoch AI, a research firm, reckons the stock of high-quality text available for training may be exhausted by 2026. Some AI labs are hiring people to write chunks of text and lines of code that models can be trained on. Others are buying synthetic data, created using computer algorithms, and hiring humans to verify it. 'Synthetic data still needs to be good data," says Wendy Gonzalez, the boss of Sama, which has operations east Africa. The other role for workers is in evaluating the output from models and helping to hammer it into shape. That is what got ChatGPT to perform better than previous chatbots. Xiaote Zhu at Scale AI provides an example of the sort of open-ended tasks being done on the firm's Outlier platform, which was launched in 2023 to facilitate the training of AI by experts. Workers are presented with two responses from a chatbot recommending an itinerary for a holiday to the Maldives. They need to select which response they prefer, rate it, explain why the answer is good or bad and then rewrite the response to improve it. Ms Zhu's example is a fairly anodyne one. Yet human feedback is also crucial to making sure AI is safe and ethical. In a document that was published after the launch of ChatGPT in 2022, OpenAI said it had hired experts to 'qualitatively probe, adversarially test and generally provide feedback" on its models. At the end of that process the model refused to respond to certain prompts, such as requests to write social-media content aimed at persuading people to join al-Qaeda, a terrorist group. Flying the nest If AI developers had their way they would not need this sort of human input at all. Studies suggest that as much as 80% of the time that goes into the development of AI is spent on data work. Naveen Rao at Databricks, an AI firm, says he would like models to teach themselves, just as he would like his own children to do. 'I want to build self-efficacious humans," he says. 'I want them to have their own curiosity and figure out how to solve problems. I don't want to spoon-feed them every step of the way." There is a lot of excitement about unsupervised learning, which involves feeding models unlabelled data, and reinforcement learning, which uses trial and error to improve decision-making. AI firms, including Google DeepMind, have trained machines to win at games like Go and chess by playing millions of contests against themselves and tracking which strategies work, without any human input at all. But that self-taught approach doesn't work outside the realms of maths and science, at least for the moment. Tech nerds everywhere have been blown away by how cheap and efficient DeepSeek's model is. But they are less impressed by DeepSeek's attempt to train AI using feedback generated by computers rather than humans. The model struggled to answer open-ended questions, producing gobbledygook in a mixture of languages. 'The difference is that with Go and chess the desired outcome is crystal clear: win the game," says Phelim Bradley, co-founder of Prolific, another AI-data firm. 'Large language models are more complex and far-reaching, so humans are going to remain in the loop for a long time." Mr Bradley, like many techies, reckons that more people will need to get involved in training AI, not fewer. Diversity in the workforce matters. When ChatGPT was released a few years ago, people noticed that it overused the word 'delve". The word became seen as 'AI-ese", a telltale sign that the text was written by a bot. In fact, annotators in Africa had been hired to train the model and the word 'delve" is more commonly used in African English than it is in American or British English. In the same way as workers' skills and knowledge are transferred to models, their vocabulary is, too. As it turns out, it takes more than just a village to raise a child. Clarification: This article has been amended to reflect Scale AI's claim that most of its labour is based in America and Europe.
Yahoo
24-03-2025
- Business
- Yahoo
DeepSeek Gets Endorsement from Apple CEO as AI Spotlight Shifts
Apple (NASDAQ:AAPL) Chief Executive Officer Tim Cook praised Chinese artificial intelligence firm DeepSeek (DEEPSEEK) during his ongoing visit to China, according to a report from the South China Morning Post. Speaking to China News Service, Cook described DeepSeek's AI models as excellent, though he did not provide additional details. DeepSeek gained attention in January with the launch of its R1 large language model, which initially sparked concern among investors about potential shifts in artificial intelligence spending. However, sentiment has since shifted. Nvidia (NASDAQ:NVDA) CEO Jensen Huang and other industry figures have said the early market reaction was likely overstated. This marks Cook's first official trip to China in 2025, as shared on his verified Weibo account. Over the weekend, Cook visited educational institutions, highlighting how Apple devices like the iPad, Mac, and Vision Pro are being used in classrooms and creative spaces. Separately, Apple on Monday said it would establish a $99 million investment fund to support clean energy development in China. This article first appeared on GuruFocus.
Yahoo
15-03-2025
- Business
- Yahoo
After DeepSeek, Chinese fund managers beat High-Flyer's path to AI
By Samuel Shen and Vidya Ranganathan SHANGHAI/SINGAPORE (Reuters) - Chinese hedge fund High-Flyer's use of artificial intelligence in trading markets has spurred an AI arms race among mainland asset managers that could shake up the country's $10 trillion fund management industry. Quant fund High-Flyer not only deployed AI in its multi-billion dollar portfolio, it also built China's most notable AI start-up DeepSeek whose cost-effective large language model stunned Silicon Valley and undermined Western dominance of the AI sector. In its wake, aspiring Chinese hedge fund managers such as Baiont Quant, Wizard Quant and Mingshi Investment Management are stepping up AI research, while dozens of mutual fund companies are rushing to incorporate DeepSeek into their investment workflow. "We are in the eye of the storm" of an AI revolution, said Feng Ji, chief executive of Baiont Quant, which uses machine learning to trade markets with no human intervention. "Two years ago, many fund managers would look at us AI-powered quants with mockery or disbelief," said Feng. "Today, these sceptics could be out of business if they don't embrace AI." Most of these funds use AI to process market data and generate trading signals based on their investors' risk profiles, rather than produce DeepSeek-like models. And as more home-grown versions of U.S. systematic trading firms such as Renaissance Technologies and are born, fund managers expect competition for "alpha", or outperformance, to intensify. Wizard Quant advertised last month to recruit top AI researchers and engineers for a lab to "reshape the future of science and technology". Demand for coding talent is heating up. Mingshi said its Genesis AI Lab is hiring computer scientists to support research and investment. In a recent roadshow, asset manager UBI Quant told investors it had already set up an AI lab several years ago to explore the use of AI in investment and elsewhere. The race to generate better trading strategies using AI requires huge computing power and high-performance chips, and local authorities said they are ready to help. For example, the government of the southern city of Shenzhen has vowed to raise 4.5 billion yuan ($620.75 million) to subsidise hedge funds' consumption of computing power, in support of their AI development. DEEPSEEK SCRAMBLE China's mutual fund industry is also scrambling to embrace AI. More than a score of retail fund companies, including China Merchants Fund, E Fund and Dacheng Fund, have completed local deployment of DeepSeek. The open-sourced, low-cost large language model has "greatly lowered the bar for AI applications" for the mutual fund industry, said Hu Yi, vice general manager of intelligent equity investment at Zheshang Fund Management. Zheshang Fund has embedded DeepSeek into its AI platform and is developing AI agents to boost efficiency of research and investment. For example, AI agents will do most of the work of junior analysts today, such as monitoring market signals and writing daily comments, "forcing humans to do more creative things," Hu said. "Before DeepSeek, AI had mostly been the realm of top tier players given the cost, talent, and technology required" but DeepSeek had "levelled the playing ground for Chinese fund managers, which are smaller than their U.S. counterparts," said Larry Cao, Principal Analyst at FinAI Research. Baiont's Feng said AI's rapid advancement offers late comers to investment management the opportunity to challenge bigger incumbents. "A seasoned fund manager may have accumulated 20 years of experience, but with AI, one can acquire that experience in two months using 1,000 GPUs," said Feng, whose five-year-old fund company currently manages 6 billion yuan, eclipsing many older rivals. ($1 = 7.2493 Chinese yuan renminbi) Sign in to access your portfolio
Iraqi News
17-02-2025
- Business
- Iraqi News
Chatbot vs national security? Why DeepSeek is raising concerns
Seoul – Chinese AI chatbot DeepSeek upended the global industry and wiped billions off US tech stocks when it unveiled its R1 programme, which it claims was built on cheap, less sophisticated Nvidia semiconductors. But governments from Rome to Seoul are cracking down on the user-friendly Chinese app, saying they need to prevent potential leaks of sensitive information through generative AI services. AFP takes a look at what's going on: Who has banned DeepSeek? First to act was Italy, which launched an investigation into DeepSeek and said it was blocking the upstart Chinese app from possessing Italian users' data. Italy's Data Protection Authority had briefly blocked Western competitor ChatGPT in 2023. Next, Taiwan banned workers in the public sector and at key infrastructure facilities from using DeepSeek, saying it was a Chinese product and could endanger national security. Australia following suit days after. Then, South Korean ministries — including defence and unification, which oversees ties with the nuclear-armed North — and the country's police force banned the app from military and work computers, citing security risks. On Monday, authorities there said that DeepSeek would not be available from local app stores while a review of its handling of personal data is carried out. US lawmakers have also moved to introduce a 'No DeepSeek on Government Devices Act', with Congressman Darin LaHood saying the national security threat that 'Chinese Communist Party-affiliated company' DeepSeek posed to the United States was 'alarming'. State-level bans were also issued in Texas, Virginia and New York. Texas Governor Greg Abbott said personal information 'must be protected from malicious espionage operations by the Chinese Communist Party.' Why are they worried? In the terms and conditions of DeepSeek, there is a section on the provision of personal data to third parties — very similar to that used by OpenAI's Chat GPT. But while US companies typically resist government requests for data, 'in China when the government requests access, companies are legally obligated to provide user data', said Youm Heung-youl, a data security professor at Soonchunhyang University. 'This distinction between respecting user privacy and providing government access often shapes how countries perceive trust in companies.' According to DeepSeek's privacy policy, it also collects information on 'key stroke patterns or rhythms' which detects how an individual interacts with each button. Is this justified? DeepSeek 'have a policy of aligning with the core values of socialism' Isabel Hou, Taiwanese AI expert and secretary-general of Taiwan AI Academy told AFP. For example, sensitive enquiries about Tiananmen Square or Taiwanese statehood –- which would typically be censored in China –- should be possible on DeepSeek elsewhere. 'But we find that DeepSeek actually uses the same set of rules when providing services overseas,' Hou added. Beijing, for its part, claims the restrictions do not reflect legitimate national security concerns but highlight 'the politicisation of economic, trade and technological issues'. It says the Chinese government 'will never require enterprises or individuals to illegally collect or store data'. Is this unexpected? 'DeepSeek was launched in May of 2023, and something like this can't just emerge overnight,' Park Seung-chan, Chinese studies professor at Yongin University told AFP. Experts point to the enormous amount of research and development (R&D) China has poured into companies in recent years. According to data from the Korea Chamber of Commerce, China ranked second among the world's top R&D investors, following the US, but showed the most significant growth, with its investment volume soaring more than 11-fold over the past decade. 'I see this (the release of R1) as a calculated move that was prepared before the Trump era, and we should pay attention to the second and third waves of DeepSeek,' said Park. What next? DeepSeek says it uses less-advanced H800 chips — permitted for sale to China until 2023 under US export controls — to power its large learning model. While semiconductor exporting powerhouses South Korea and Taiwan have been thriving on sales of cutting-edge chips, DeepSeek has thrown the industry into turmoil. 'If DeepSeek really used H800, it means that even without cutting-edge semiconductors, similar outcomes could be achieved with general semiconductors, as long as the software is good,' Park Ki-soon, a professor of Chinese economics at Sungkyunkwan University told AFP. 'Countries like the US and China are investing massive amounts of talent and resources into software development,' he said, adding that DeepSeek showed governments needed to boost this further and 'provide support to foster this growth'.
Associated Press
13-02-2025
- Business
- Associated Press
BlipCut Integrates DeepSeek Model to Enhance AI Video Translation Quality
NEW YORK, N.Y., Feb. 13, 2025 (SEND2PRESS NEWSWIRE) — BlipCut, a leading innovator in video translation and localization, is excited to announce its integration of the DeepSeek model to enhance translation rewriting capabilities. The integration of DeepSeek enables BlipCut AI Video Translator to refine translated text with an advanced AI-powered approach, ensuring that subtitles and voiceovers better align with natural language nuances. This enhancement empowers content creators, businesses, and media platforms to reach global audiences with higher-quality localized content. KEY IMPROVEMENTS BROUGHT BY DEEPSEEK INTEGRATION INCLUDE: Better Audio-Video Synchronization: Speech speed is adjusted to align seamlessly with on-screen visuals, ensuring a more immersive viewing experience. Higher Translation Accuracy: Improved contextual understanding results in more precise translations, preserving the original meaning and tone. DeepSeek is particularly effective for technical and specialized content, ensuring terminology is accurate and industry-specific wording is used. It focuses on technical features that make translations more reliable in areas such as engineering, medicine and software development. Faster Translation Processing: With DeepSeek's advanced AI capabilities, video translations are now completed significantly faster, reducing wait times and streamlining the content creation process. Enhanced Translation Quality: DeepSeek improves the clarity, readability, and fluency of translations by refining sentence structure and word choices. This ensures subtitles and voiceovers are more natural, contextually appropriate, and engaging for audiences. BlipCut's upgraded translation system not only improves linguistic accuracy, but also maintains the intent and tone of the source files. This is particularly beneficial for industries such as entertainment, e-learning and corporate communications, where accurate translations are critical to audience engagement and understanding. BlipCut users now have access to enhanced translation capabilities powered by DeepSeek. As AI technology continues to evolve, BlipCut remains committed to advancing video localization and ensuring that global storytelling is more inclusive and accessible than ever before. About BlipCut: BlipCut is an industry-leading AI video translation platform that allows users to break language barriers with ease. With features like batch translator, subtitle generation, AI voiceover, ai clip generator, and support for 130+ languages, BlipCut is the go-to solution for content creators, educators, and businesses worldwide. NEWS SOURCE: BlipCut Keywords: Artificial Intelligence, BlipCut deepseek translation, AI video tooks, apps, software, Video Translation, NEW YORK, N.Y. Send2Press® Newswire. Information is believed accurate but not guaranteed. Story ID: S2P124053 AP-R15TBLLI RIGHTS GRANTED FOR REPRODUCTION IN WHOLE OR IN PART BY ANY LEGITIMATE MEDIA OUTLET - SUCH AS NEWSPAPER, BROADCAST OR TRADE PERIODICAL. MAY NOT BE USED ON ANY NON-MEDIA WEBSITE PROMOTING PR OR MARKETING SERVICES OR CONTENT DEVELOPMENT.
