CBNFridaySpecial丨WhatisChatGPTcrazeallabout？

　　Hello! Welcome to this edition of CBN Friday Special. I’m Stephanie Li.
　　A logical thinker, smart, savvy, sophisticated, sometimes even with a touch of humor.
　　No, I’m not describing my best friend, or any human being at all, but rather, talking about ChatGPT, who is, in essence, a simple online artificial intelligence chatbot created by the San Francisco company OpenAI in December 2022.
　　According to Elon Musk, who was one of the founders of OpenAI before leaving, ＂ChatGPT is scary good. We are not far from dangerously strong AI.＂
　　So, if you haven’t caught up with the latest AI fad about ChatGPT, here’s how it works. You can ask it a question, and it will give the answer to that question. In a blink.
　　If you ask ChatGPT what it is, this is what it will tell you: I’m an AI-language model developed by OpenAI and capable of generating human-like text based on given input.
　　The technology that underlies ChatGPT is referenced in the second half of its name, GPT, which stands for Generative Pre-trained Transformer. Transformers are specialized algorithms for finding long-range patterns in sequences of data. A transformer learns to predict not just the next word in a sentence but also the next sentence in a paragraph and the next paragraph in an essay. This is what allows it to stay on topic for long stretches of text.
　　Because a transformer requires a massive amount of data, it is trained in two stages: first, it is pre-trained on generic data, which is easier to gather in large volumes, and then it is fine-tuned on tailored data for the specific task it is meant to perform. ChatGPT was pre-trained on a vast repository of online text to learn the rules and structure of language; it was fine-tuned on dialogue transcripts to learn the characteristics of a conversation.
　　And because of the stir its advanced capabilities have made, a lot of people are flocking to use it. ChatGPT has got mega-popular in just a matter of weeks—way faster than social media platforms like TikTok or Instagram.
　　Only two months after its launch in late November, the chatbot had 100 million monthly active users in January. OpenAI, which owns and hosts ChatGPT, recently became one of the 50 most visited websites in the world. For context, it took Instagram two and a half years to get to 100 million. TikTok got there in nine months.
　　Known as a generative AI, it responds to virtually any prompt you give it with startling speed and clarity. Whereas many chatbots only know how to respond to certain keywords or triggers, ChatGPT can respond to complex questions and spit out comprehensive, essay-length answers on virtually any topic.
　　Though unavailable yet on the Chinese mainland, Chinese internet users are finding creative uses for ChatGPT. A Chinese programmer shared his experience of using ChatGPT to write code in Visual Basic to automatically manipulate Excel files on a video he uploaded. ＂I got the code I wanted by simply describing my needs, it greatly improved my work efficiency,＂ he said.
　　Chinese technology companies and computer scientists are closely monitoring public interest in the field of artificial intelligence generated content (AIGC), and following closely behind.
　　＂ChatGPT seems to open a window to artificial general intelligence. Foundation models tend to have the ability of reasoning, induction and understanding of common sense,＂ said a computer science professor, adding that the technology was now showing the ability to develop ＂semantic understanding＂.
　　ChatGPT is able to do this by running the Internet’s vast amounts of data through powerful neural networks: software loosely designed on neurons in the human brain. However, the technology is not without flaws. ChatGPT answers users’ questions by learning from massive amounts of text on the internet, and making guesses based on probability. However, unlike a search engine, it still cannot answer questions based on real-time information and sometimes the answers it gives will sound plausible but will be incorrect.
　　It has also been ＂trained＂ using data from 2021 and before, so cannot answer questions about things that happened in the last two years.
　　In fact, despite ChatGPT’s astonishing rise to fame recently, its peers, Chatbots, have been around for decades, and you might know them best from interacting with any customer service agent over text; often, that＂s a chatbot.
　　All chatbots do not use AI like ChatGPT does. Some chatbots simply use keywords to create a response that may or may not be helpful. For instance, if you tell a non-AI chatbot that you didn＂t receive a package that was marked as delivered, it might encourage you to check the front door. Other chatbots, like the ChatGPT, use some pretty sophisticated AI.
　　The intelligent language tool – which can write cover letters, produce dating app messages, and even offer up generic investing advice – has become a buzzword in the world of finance, while some venture capitalists predict it＂s the ＂next platform shift＂ in tech.
　　More and more big tech names in the world are throwing their hats in the ring and try to get a head start before competitors get in the blue sea.
　　Microsoft said in January that it would pour $10 billion into ChatGPT creator OpenAI as part of a ＂multiyear, multibillion-dollar＂ investment. It had already provided $1 billion in backing in 2019.
　　The tech giant on Thursday released a version of Teams premium that uses OpenAI technology, and it＂s reportedly planning to launch a version of its search engine Bing powered by ChatGPT. Getting ChatGPT into the little-used Bing is likely to give Microsoft a considerable edge over its Big Tech rivals and lift its share price, analysts have said.
　　Google parent Alphabet is apparently under threat from the rise of ChatGPT, because the AI bot is a potential rival to its search engine. In December, the company＂s management issued a ＂code red＂ about the bot＂s launch. Google debuted a potential rivals to OpenAI＂s tech on Monday, a homegrown chatbot called ＂Bard＂.
　　However, it turned out to be an epic flop. The rollout of Google’s highly anticipated chatbot resulted in a $100 billion fumble on Wednesday after the Bard spit out inaccurate information in a company ad, about which satellite first took pictures of a planet outside the Earth＂s solar system.
　　It＂s not just US big techs. Chinese tech companies are upping the ante in the fast-growing AIGC sector. Baidu, Tencent and Alibaba have invested heavily in promoting the commercial use of AI, and some Chinese AI unicorns have grown rapidly in recent years.
　　Baidu Inc announced on Tuesday that it will complete internal testing of its AI chatbot service similar to ChatGPT, called ＂Ernie Bot＂ in March.
　　The Beijing-based company has invested large sums of money in developing its Ernie system, a large-scale machine-learning model that has been trained on massive data over several years and possesses in-depth semantic comprehension and generation capabilities.
　　Alibaba said it is developing a ChatGPT style AI tool and that it was currently in internal testing, while gaming major NetEase plans to deploy similar ＂large language model＂ technology in its education business. 360 Security Technology, in response to regulators＂ queries, said its self-developed ChatGPT-related technology is still at a nascent stage and is used only internally as a productivity tool.
　　ChatGPT has taken the stock market by storm, massively boosting the value of companies who say they＂ll incorporate similar technology into their business.
　　AIGC-related stocks continued to rally in the Chinese A-share market, with Chinese AI companies, such as Cloudwalk Technology and Speechocean, seeing their shares surged by the daily limit for days within this week.
　　Chinese state media on Thursday cautioned against risks in chasing local ChatGPT-concept stocks, while domestic AI companies urged investors to be rational after their soaring share prices caught regulators＂ attention.
　　In a front-page editorial, the Securities Times highlighted several technological concepts that previously spurred stock buying in China - such as 5G, AR, VR and anti-virus garments - the excitement for which has died down.
　　Though some hotly chased concepts have been successful, ＂many more new ideas haven＂t been commercialized, or require more time to prove,＂ the state-backed newspaper said. Companies developing ChatGPT-like concepts have also flagged risks at the request of regulators after their prices shot up amid intense interest in generative AI technology.
　　Regardless, the sensational popularity of ChatGPT reveals both its usefulness in helping with a wide range of tasks and a general overflowing curiosity about human-like machines. Experts are split on whether this marks the dawn of a new AI-era or if the hype will die down as people hit the limits of ChatGPT’s current capabilities.
　　ChatGPT is already causing problems.
　　Perhaps unsurprisingly, people have already found uses for the tool that are less than ideal. For example, someone could use the app to create code for malware, students will use the speedy text generator to cheat or plagiarize, and since the chatbot isn’t always accurate: its sources aren’t fact-checked, it might be used to spread disinformation.
　　And the biggest concern is probably an existential crisis for content creators — writers, screenwriters, journalists, coders, are among the jobs that might get seriously threaten by AIGC technology.
　　As with every wave of automation technologies, the latest will likely have a significant impact on jobs and the future of work. Whereas blue-collar workers bore the brunt of earlier waves, generative AI will likely have a greater effect on white-collar professions. A 2019 study from the Brookings Institution found that AI would most affect jobs such as marketing specialists, financial advisers and computer programmers.
　　Those effects will be mixed. Economists who study automation have found that three things tend to happen: some workers improve their productivity, some jobs are automated or consolidated, and new jobs that didn’t previously exist are also created.
　　The final scorecard is difficult to predict. In company-level studies of automation, researchers have found that some companies that adopt automation may increase their productivity and ultimately hire more workers over time. But those workers can experience wage deflation and fewer career growth opportunities.
　　As OpenAI explains, the tool will likely yield a lot of false positives and negatives, sometimes with great confidence. ＂ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness,＂ tweeted OpenAI Chief Executive Sam Altman shortly after the chatbot’s release, adding that it is ＂a mistake to be relying on it for anything important right now.＂
　　＂As a language model, I can not predict the future. It is important to note that while some jobs may be lost due to AI, new job opportunities are also likely to be created as businesses and organizations adopt and make use of these new technologies. The key challenge will be to ensure that workers have the skills and training needed to transition into these new job opportunities.＂
　　And that’s part of the answer ChatGPT gave to the question ＂which jobs do you think ChatGPT will replace＂.
　　忽然之间，朋友圈都在讨论一款名叫ChatGPT的AI程序。
　　ChatGPT是一款聊天机器人模型ChatGPT，能够模拟人类的语言行为，与用户进行自然的交互。ChatGPT于2022年11月发布，开发者是微软投资的人工智慧研究实验室Open AI。
　　一经问世，ChatGPT就被用户们疯狂＂调戏＂，有人用其写小作文，有人拿高考题来考验它，有人让它写代码，甚至一位美国的大学教授尝试让它考MBA，结果还考了高分……随后，ChatGPT越来越多功能被挖掘出来，不仅＂文能写文章、武能改Bug＂，被马斯克评价为＂厉害得吓人＂，更因为＂未来是否有可能取代搜索引擎＂而引发热议。
　　ChatGPT上线短短5天，Open AI首席执行官Sam Altman就在社交媒体上披露，用户已经突破了100万。基于GPT-3模型开发，ChatGPT使用大量训练数据模拟人类语言行为，通过语义分析生成文本从而与用户进行自然交互。公开数据显示，截至2023年1月末，ChatGPT的月活用户已经突破1亿，成为史上用户增长速度最快的消费级应用程序。
　　之所以这么火，在于ChatGPT不仅能跟人类聊天逗闷子，还可能改变人类未来的生活和工作。各个领域的文字工作者、工程师、学生、教授等，可运用ChatGPT自动生成文字，摘要，表格，翻译。此外，它还具有编写和调试计算机程序，创作音乐、电影电视剧剧本、歌词、各学科论文、童话故事、小说的能力。
　　ChatGPT操作简单，容易上手，打开之后在对话框里输入问题就能获得答案。而且，它就像一个全能选手，常常能够给人出乎意料的答案。应对日常对话，ChatGPT不在话下，可以分辨一些提问中的描述错误，面对不合理、不道德的请求时，能直接拒绝。
　　ChatGPT令人惊喜的一个表现在于，能够进行文学创作。比如，给ChatGPT一个话题，它就可以写小说框架。有用户体验之后评价称，＂ChatGPT的语言组织能力、文本水平、逻辑能力，可以说已经令人惊艳了＂。甚至已经有用户打算把日报、周报、总结反思这些文字工作，都交给ChatGPT来辅助完成。
　　普通的文本创作，只是最基本的，ChatGPT还能给程序员的代码找Bug，一些开发者在试用中表示，ChatGPT针对他们的技术问题提供了非常详细的解决方案，比一些搜索软件的回答还要靠谱。美国代码托管平台Replit首席执行官Amjad Masad在社交媒体发文称，ChatGPT是一个优秀的＂调试伙伴＂，＂它不仅解释了错误，而且能够修复错误，并解释修复方法＂。
　　ChatGPT更具备商业逻辑，不仅非常了解自己的优劣势，可以为自己进行竞品分析、撰写营销报告，就连世界经济形势也＂了如指掌＂，能说出自己的见解。
　　有人将ChatGPT比喻为＂搜索引擎+社交软件＂的结合体，能够在实时互动的过程中获得问题的答案。国盛证券区块链研究院在研报中表示，ChatGPT在寻找答案、解决问题的效率上已经部分的超越了如今的搜索引擎，ChatGPT或许在未来会改变我们获取信息、输出内容的方式，AIGC有望成为数字经济时代驱动需求爆发的杀手级应用。
　　ChatGPT能够实现当前的交互，离不开OpenAI在AI预训练大模型领域的积累。OpenAI是微软投资的AI实验室，也是全球AI领域最为领先的AI实验室之一，一直因为技术层面的领先，被当作是AI领域的行业＂灯塔＂。
　　从2018年起，OpenAI就开始发布生成式预训练语言模型GPT（Generative Pre-trained Transformer），可用于生成文章、代码、机器翻译、问答等各类内容。在当时，参数量只有1.17亿。
　　接下来，OpenAI每一代GPT模型的参数量都爆炸式增长，2019年2月发布的GPT-2参数量为15亿，而2020年5月的GPT-3，参数量达到了1750亿，直接带动了预训练大模型成为AI领域近两年的发展趋势，掀起一场拼参数、拼算力的军备竞赛。GPT-4还未到来，这一次发布的ChatGPT，被视为是基于GPT-3的微调版本，即GPT-3.5。
　　OpenAI表示，ChatGPT采用的模型使用了＂利用人类反馈强化学习（RLHF）＂的训练方式，包括了：人类提问机器答、机器提问人类回答，并且不断迭代，让模型逐渐有了对生成答案的评判能力。相比GPT-3，ChatGPT的主要提升点在于记忆能力，可实现连续对话，极大地提升了对话交互模式下的用户体验。
　　但是，ChatGPT也并不是总能给人带来惊喜，不可否认，它还存在许多局限性，尤其对于一些文化常识问题、数学计算题不是太擅长。而且，ChatGPT的回答往往是大段长段，过于冗长，看似逻辑自洽，但有时是在一本正经地＂忽悠人＂，仔细思索会发现存在一定漏洞。如果非专业人士无法分辨ChatGPT答案的准确性，极有可能会被严重误导。
　　即便好评和质疑齐飞，也挡不住一大波ChatgptGPT的赶潮儿纷沓而至。ChatGPT出圈了，也让其所处的AIGC板块关注度飙升。
　　所谓AIGC，即利用人工智能技术来生成内容。2022年，Web 3.0时代加速到来，对于内容生产的数量和质量有更高的需求。和此前Web1.0、Web2.0时代的UGC（用户生产内容）和PGC（专业生产内容）相比，代表人工智能生产内容的AIGC，是新一轮内容生产方式的变革，AIGC内容在Web3.0时代也将出现指数级增长。
　　百度创始人李彦宏就曾表示，＂AIGC或许将颠覆现有内容生产模式，实现以‘十分之一的成本’，以百倍千倍的生产速度，去生成AI原创内容。＂
　　自去年年底以来大热的ChatGPT，也在迎来越来越多的竞争者。
　　当地时间2月6日，谷歌公司首席执行官桑达尔·皮查伊在一篇博文中宣布了ChatGPT竞争对手Bard，将其描述为＂实验性的对话式人工智能服务＂。然而让人大跌眼镜的是，被谷歌寄予厚望的Bard在首发推广的广告里，却因为回答了一个天文问题就被指出犯了事实错误。Bard首战出师不利，连累谷歌周三股价大跌，市值蒸发上千亿美元。
　　微软更是紧锣密鼓地提速相关应用步伐。近日，微软宣布将ChatGPT整合至全线产品中，并扩大与ChatGPT所有者OpenAI的合作关系。集成了ChatGPT-4的新版Bing（必应）亦曾短暂上线，部分用户提前访问了新版Bing，体验到以聊天方式展示搜索结果的全新搜索引擎。
　　2月7日，百度透露内部正在进行ChatGPT竞品的上线冲刺，该项目名字确定为文心一言，英文名ERNIE Bot。项目将于3月份完成内测，并面向公众开放。消息一出，百度股价再度大涨。
　　事实上，百度在ChatGPT所在的AIGC领域已有布局。早在去年9月，百度CEO李彦宏就判断人工智能发展在＂技术层面和商业应用层面，都有方向性改变＂。12月，在一场面向全体员工的内部直播中，李彦宏更是直接提及AIGC（人工智能技术生成内容）与ChatGPT。
　　除百度之外，国内厂商也在试图抓住ChatGPT带来的热度。2月7日，三六零在互动平台表示，公司的人工智能研究院从2020年开始一直在包括类ChatGPT技术在内的AIGC技术上有持续性的投入。
　　不过，三六零也坦言，截至目前其相关技术仅作为内部业务自用的生产力工具使用，且投资规模及技术水平与当前的ChatGPT 3相比还有较大差距，各项技术指标只能做到略强于ChatGPT 2。公司也计划尽快推出类ChatGPT技术的demo版产品。
　　＂尽管当前掀起ChatGPT的热潮，但我们认为短期内ChatGPT对市场影响有限，长期来看这些模型甚至在3年以后会退出市场。＂在接受21世纪经济报道记者采访时，IDC中国研究总监卢言霞直言道，＂基于ChatGPT可能会诞生一批新创企，但单纯依赖大模型无法为创企提供持续的竞争力。垂直领域的数据、面向场景的模型优化、工程化的解决方案，才是将AI落地的根本，也是建立竞争优势的关键。＂
　　关于ChatGPT对人类工作的挑战，也引发了广泛的忧虑。
　　很多用户抱着好玩的心态和ChatGPT互动，但伴随着体验的深入，也会担心，自己会不会有朝一日因为AI而失业。尽管ChatGPT并未彻底改变我们现在的生产方式，但它却证明了人工智能已发展至新的阶段。
　　那么，ChatGPT自己是如何看待人工智能可能取代人类工作的问题呢？
　　＂人工智能对就业市场的影响是一个复杂的问题，专家们对此进行了广泛争论。虽然人工智能确实有可能使某些任务自动化，但也有可能在人工智能开发和维护、数据分析和技术管理等领域创造新的就业机会。值得注意的是，由于自动化导致的工作岗位流失并不是一个新现象，并且随着新技术的出现已发生了几个世纪。关键是要确保工人具备适应这些变化和利用新机会所需的技能。＂
　　Executive Editor: Sonia YU
　　Editor: LI Yanxia
　　Host: Stephanie LI
　　Writer: Stephanie LI
　　Sound Editor: Stephanie LI
　　Graphic Designer: ZHENG Wenjing, LIAO Yuanni
　　Produced by 21st Century Business Herald Dept. of Overseas News.
　　Presented by SFC
　　编委: 于晓娜
　　策划、编辑：李艳霞
　　播音：李莹亮
　　撰稿：李莹亮
　　音频制作：李莹亮
　　设计：郑文静、廖苑妮
　　21世纪经济报道海外部 制作
　　南方财经全媒体集团 出品
　　更多内容请下载21财经APP