跟读练习: What is AI governance? - 通过视频学习英语口语
选择练习模式
每种模式都会打开全屏练习室。
34 句
如果句子过短或过长，请点击 Edit 进行调整。
1
Hello, my name is Kush Varshney.
2
I'm a distinguished research scientist and senior manager at IBM Research.
3
I'm here to answer three questions about AI governance.
4
Governance is an interesting word.
5
It means control.
6
And if we think back in history, steam engines had this part called the governor and it regulated the flow of steam
7
and it was there to make sure that the system remains safe.
8
And when we say AI governance, it's the set of practices to keep an AI system under control so that it remains safe as well.
9
So it includes things like respecting regulations, curating data, and creating fact sheets that explain how the system works.
10
Some of the risks posed by generative AI and LLMs are actually the same as traditional machine learning.
11
So things like fairness, transparency, robustness to attacks, they're all still concerns.
12
But then there's some new ones as well.
13
So new risks that are emerging with generative AI include things like toxicity, harmful behaviors, bullying, these sort of things.
14
And then there's also hallucination.
15
So where the model is putting out things which kind of sound like they would make sense, they're plausible, but they're actually factually incorrect.
16
And we've created a taxonomy of common harms that can mitigate these behaviors with data curation, with fine tuning, prompt tuning, prompt engineering.
17
So all of that is really the way to address them.
18
But this is just a starting point.
19
So if we start and think for a second,
20
governments, enterprises, all sorts of other organizations will have additional constraints to build into their LLMs so that they follow applicable laws,
21
industry standards, social norms, and so on.
22
And understanding the models is also quite different
23
because we no longer need to know exactly how the models make their decisions and predictions and predictions.
24
This is what we used to call explainability, but now instead what we need to be able to do
25
is trace an LLM's generative output to the user's prompt or to its training data.
26
So source attribution is the new explainability.
27
Safety guarantees with generative AI are very difficult because there's no one-size-fits-all approach.
28
Every organization needs to define their own values, their own behaviors that they find acceptable and unacceptable.
29
For example, once you know the context of your chatbot's deployment, you can start to reason through the relevant risks.
30
This could involve taking existing policy documents like laws,
31
corporate policies, or other rules and using them as instruction data to teach the model how to behave.
32
And interventions must also be scalable because we're dealing with such large models and even larger data sets.
33
And at IBM we're designing tools to address the unique threats posed by large language models and other content generating foundation models.
34
We're also devising new methods to limit chatbots from leaking personal or proprietary data and hallucinating wrong or irrelevant information.
同频道
FOR YOU
推荐视频
C12
250 句
AI Strategy, Policy, and Governance | Allan Dafoe
Business EnglishUniversity StudentsWorking Adults
C111
98 句
What are AI Agents?
Business EnglishUniversity StudentsWorking Adults
C13
208 句
AI & Education: Generative AI & the Future of Critical Thinking
Business EnglishUniversity StudentsWorking Adults
C143
6 句
Ray Dalio on the AI Bubble
Business EnglishUniversity StudentsWorking Adults
C11
83 句
Leading AI CEOs Warn UN Security Council of Existential Risks | Sam Altman On AI Controls | N18G
Business EnglishDaily ConversationUniversity Students
C11
402 句
FULL DISCUSSION: Google's Demis Hassabis, Anthropic's Dario Amodei Debate the World After AGI | AI1G
IELTS PrepBusiness EnglishUniversity Students
C13
219 句
Why AI is Legal’s New Competitive Advantage - Future of Work Podcast
Business EnglishUniversity StudentsWorking Adults
C11
127 句
The Alignment Problem Explained: Crash Course Futures of AI #4
Business EnglishDaily ConversationUniversity Students
![What is AI and ML? [Pt 2] | Generative AI with .NET for Beginners](https://img.youtube.com/vi/Len8KRBFvGo/mqdefault.jpg)
B22
52 句
What is AI and ML? [Pt 2] | Generative AI with .NET for Beginners
Business EnglishUniversity StudentsWorking Adults
C11
105 句
AI Governance Explained: Ethical, Transparent & Accountable AI for the Future
Daily ConversationUniversity StudentsWorking Adults
C11
114 句
Harnessing generative AI in architecture, engineering & construction
Daily ConversationUniversity StudentsWorking Adults
真实场景：解读AI治理
视频中，IBM研究人员以通俗易懂的方式解释“AI治理”，从历史上的蒸汽机调速器切入，对比传统机器学习与生成式AI的风险，还提到企业如何通过数据管理、模型调优应对挑战。这样的科技对话场景在学术讨论、职场汇报中很常见，既能学习专业表达，也能锻炼听懂复杂概念的能力。
实用表达与搭配
- keep...under control：使……处于控制之中。例：AI governance aims to keep AI systems under control.（AI治理旨在让AI系统处于可控状态。）
- pose risks：带来风险。例：Generative AI poses new risks like hallucination.（生成式AI带来了幻觉等新风险。）
- trace...to：追溯……到。例：We need to trace the output to training data.（我们需要将输出追溯到训练数据。）
- one-size-fits-all approach：一刀切的方法。例：Safety guarantees need no one-size-fits-all approach.（安全保障不需要一刀切的方法。）
影子跟读挑战：提升口语与理解
影子跟读（shadowspeak）是提高英语发音和流畅度的有效方法。现在就试试：播放视频中“Governance is an interesting word...so that it remains safe as well”这段，暂停后重复，注意模仿重音（如“control”“safe”）和语调。完成后，用“keep...under control”造句，描述一个生活场景（如“Parents try to keep their kids' screen time under control”）。通过这种“看视频学英语+即时练习”的方式，既能掌握专业词汇，又能强化口语表达。记住，shadow speak的关键是“同步”与“模仿”，坚持一周，你会发现发音和听力理解都有明显进步。
什么是跟读法？
跟读法 (Shadowing) 是一种有科学依据的语言学习技巧，最初开发用于专业口译员的培训，并由多语言者Alexander Arguelles博士普及。这个方法简单而强大：您在听英语母语原声的同时立即大声重复——就像是一个延迟1-2秒紧跟说话者的影子。与被动听力或语法练习不同，跟读法强迫您的大脑和口腔肌肉同时处理并模仿真实的讲话模式。研究表明它能显着提高发音准确性，语调，节奏，连读，听力理解和口语流利度——使其成为雅思口语备考和真实英语交流最有效的方法之一。