​2023: A Big Year for Artificial Intelligence

2023: A Big Year for Artificial Intelligence

2023：人工智能的大年

【听力音频】

----------------------------------------------------------------------------------

　　By John Russell

　　20 December 2023

　　Artificial intelligence (AI) became a popular subject in 2023. Still, the technology has a long way to go to meet people's futuristic expectations of human-like machines.

　　ChatGPT was central to this year's attention on AI. The chatbot showed the world recent developments in computer science, although not everyone understood quite how it works or what to do with it.

　　AI scientist Fei-Fei Li suggested that 2023 would be remembered for the great changes in technology as well as the public awakening. It was a year for people to figure out "what this is, how to use it, what's the impact — all the good, the bad and the ugly," she added.

　　Concerns over AI

　　The first AI concerns of 2023 began soon after New Year's Day. That is when classrooms reopened and schools from Seattle to Paris started blocking ChatGPT. Students were already asking the chatbot — released in late 2022 — to write papers and answer take-home tests.

　　AI large language models behind technology such as ChatGPT work by predicting the next word in a sentence. The models make such predictions after having "learned" the structure of a huge number of human-written works. The large language models often get facts wrong. But the results appeared so natural that it created interest in the next AI developments and possible uses for trickery and deception.

　　Worries grew as this new group of generative AI tools produced not just words but also images, music and voices. They seemed to threaten the jobs of anyone who writes, draws, and creates music and computer languages. Concerns about AI tools fueled strikes by Hollywood writers and actors and legal disputes from artists to writers.

　　Some of the most respected AI scientists warned that the technology's progress was marching toward outsmarting humans and possibly threatening their existence. Yet other scientists called the warnings overblown and brought attention to more immediate risks.

　　By spring, AI-created videos known as deepfakes had appeared in U.S. election campaigns. Deepfakes are videos that contain realistic images but with digital changes to people's actions and speech. One deepfake falsely showed Donald Trump embracing the nation's former top infectious disease expert. The technology made it increasingly difficult to tell the difference between real and fake videos of the wars in Ukraine and Gaza.

　　By the end of the year, the AI crisis affected ChatGPT's maker, OpenAI. The San Francisco-based company led by chief executive Sam Altman was nearly destroyed by disagreements over its mission.

　　AI debates also led to new laws from the European Union and consideration from others, including the United States Congress.

　　Too much excitement?

　　AI products released in 2023 have brought technology achievements not possible in earlier times. But the market research company Gartner says they arrive with "inflated expectations" and "massive claims" about its abilities.

　　Gartner analyst Dave Micko said leading AI developers are pushing the latest technology with their current line of products, including search engines and workplace productivity software.

　　He said, "As much as Google and Microsoft and Amazon and Apple would love us to adopt the way that they think about their technology and that they deliver that technology, I think adoption actually comes from the bottom up."

　　It is easy to forget that this is not the first appearance of AI in business. The technology has been used to help guide self-driving cars, compare objects and individual faces, and recognize speech in software like Siri and Alexa.

　　In 2011, Tom Gruber launched Siri, which was bought and used by Apple in the iPhone. At that time, it was the only major use of AI that people had ever experienced.

　　But Gruber believes what's happening now is the "biggest wave ever" in AI, launching new possibilities as well as dangers.

　　The dangers could come fast in 2024. Major national elections in the U.S., India and elsewhere could get flooded with AI-created deepfakes.

　　In the longer term, AI technology's rapidly improving language, visual sensing and step-by-step planning abilities could create a kind of digital assistant — but only if given access to the "inner loop of our digital life stream," Gruber said.

　　"They can manage your attention as in, ‘You should watch this video. You should read this book. You should respond to this person's communication,'" Gruber said. He added, "That is what a real executive assistant does. And we could have that, but with a really big risk of personal information and privacy."

　　I'm John Russell. And I'm Ashley Thompson.

　　Matt O'Brien reported on this story for the Associated Press. John Russell adapted it for VOA Learning English.

----------------------------------------------------------------------------------

　　作者：John Russell

　　2023年12月20日

　　人工智能（AI）在2023年成为了热门话题。然而，这项技术要达到人们对于类人机器的未来期望，还有很长的路要走。

　　ChatGPT成为了今年人工智能关注的焦点。这个聊天机器人向世界展示了计算机科学的最新发展，尽管并不是每个人都完全理解它是如何工作的，或者应该如何使用它。

　　人工智能科学家李飞飞建议，2023年将因技术的巨大变化以及公众的觉醒而被记住。这是一个让人们弄清楚“这是什么，如何使用它，影响是什么——所有的好的，坏的和丑陋的”之年，她补充说。

　　对AI的担忧

　　2023年对AI的首次担忧始于新年后不久。那时，教室重新开放，从西雅图到巴黎的学校开始屏蔽ChatGPT。学生们已经开始向这个在2022年底发布的聊天机器人询问写论文和回答家庭作业的问题。

　　像ChatGPT这样的技术背后的大型AI语言模型是通过预测句子中的下一个单词来工作的。这些模型在“学习”了大量人类写作作品的结构后做出这样的预测。大型语言模型经常会出错。但结果看起来如此自然，以至于它引发了对下一步AI发展和可能用于欺骗和欺诈的用途的兴趣。

　　随着这个新的生成AI工具群不仅产生了文字，还产生了图像、音乐和声音，人们的担忧增加了。他们似乎威胁到任何写作、绘画、创作音乐和计算机语言的人的工作。对AI工具的担忧引发了好莱坞作家和演员的罢工，以及从艺术家到作家的法律纠纷。

　　一些最受尊敬的AI科学家警告说，这项技术的进步正在向超越人类智能的方向迈进，可能威胁到人类的生存。然而，其他科学家认为这些警告过于夸大，他们将注意力转向了更直接的风险。

　　到了春天，被称为深度伪造的AI创造的视频开始出现在美国的选举活动中。深度伪造是包含逼真图像的视频，但对人们的行为和言语进行了数字化的改变。一部深度伪造的视频错误地展示了唐纳德·特朗普拥抱国家前顶级传染病专家的场景。这项技术使得人们越来越难以区分乌克兰和加沙战争的真实视频和伪造视频。

　　到了年底，AI危机影响到了ChatGPT的制造商OpenAI。这家由首席执行官Sam Altman领导的位于旧金山的公司几乎因为对其使命的分歧而崩溃。

　　AI的争论也导致了欧盟的新法律，并引起了其他国家，包括美国国会的考虑。

　　太激动了吗？

　　2023年发布的AI产品带来了以往无法实现的技术成就。但市场研究公司Gartner表示，这些产品带来了对其能力的“过高期望”和“巨大声明”。

　　Gartner分析师Dave Micko表示，领先的AI开发商正在用他们当前的产品线推动最新的技术，包括搜索引擎和工作场所生产力软件。

　　他说：“尽管谷歌、微软、亚马逊和苹果都希望我们接受他们对自己技术的理解和他们提供技术的方式，但我认为实际的采用是从底层逐步推广的。”

　　人们很容易忘记，这并不是AI在商业中的首次出现。这项技术已经被用于帮助引导自动驾驶汽车，比较物体和个体面孔，以及在像Siri和Alexa这样的软件中识别语音。

　　2011年，Tom Gruber推出了Siri，后被苹果购买并用于iPhone。那时，它是人们唯一经历过的主要AI应用。

　　但Gruber认为，现在正在发生的是AI的“最大的浪潮”，带来了新的可能性以及危险。

　　2024年的危险可能会迅速到来。美国、印度和其他地方的重大国家选举可能会被AI创造的深度伪造视频淹没。

　　从长远来看，AI技术的语言、视觉感知和逐步规划能力的快速提高可能会创造出一种数字助手——但前提是要让它们能够接触到我们的“数字生活流的内环”，Gruber说。

　　“它们可以管理你的注意力，比如，‘你应该看这个视频。你应该读这个书。你应该回应这个人的沟通，’”Gruber说。他补充说，“这就是一个真正的行政助手所做的。我们可以拥有这样的助手，但是个人信息和隐私的风险非常大。”

　　我是John Russell。我是Ashley Thompson。

　　这个故事是由美联社的Matt O'Brien报道的，John Russell为VOA学习英语改编。

----------------------------------------------------------------------------------

　　本故事中的词汇

　　chatbot -- n. 聊天机器人，一种设计用来与人对话的计算机程序

　　impact -- n. 影响，强大或主要的影响或效果

　　deception -- n. 欺骗，使某人相信不真实的事情的行为

　　embrace -- v. 拥抱，用你的手臂抱住某人以表达爱或友谊

　　manage -- v. 管理，照顾并做出关于某事的决定

　　respond - v. 回应，对问题或请求说或写出某事作为答案