摘要：I'm thrilled to announce that SoftBank will be investing 100 billion dollars in America, creating 100,000 American jobs at a minim
A major announcement this week from president-elect Donald Trump and SoftBank CEO Masayoshi Son.
当选总统唐纳德·特朗普和软银首席执行官孙正义本周发布了一项重大公告。
I'm thrilled to announce that SoftBank will be investing 100 billion dollars in America, creating 100,000 American jobs at a minimum.
我很高兴地宣布，软银将在美国投资1000亿美元，至少为美国创造10万个就业岗位。
It will help ensure that artificial intelligence, emerging technologies, and other industries of tomorrow are built created and grown right here in the USA.
它将有助于确保人工智能、新兴技术、以及其他未来行业都是在美国建立、创造和发展。
Innovations in generative artificial intelligence are advancing at a rapid pace, fueled in part by the popularity of large language model programs like ChatGPT, which are trained on vast quantities of textual data to generate new content with the technology, seemingly unlimited capabilities raise questions about safety and regulation.
生成人工智能的创新正在快速推进，部分原因是ChatGPT等大型语言模型程序的盛行，它们在大量文本数据上进行训练，以使用该技术生成新内容，看似无限的能力引发了有关安全和监管的质疑。
The Trump Administration has been very clear, that it wants to eliminate a lot of the AI regulation that was put into place in the Biden Administration, saying they want to unshackle AI to win the AI war to lead the world.
特朗普政府已经非常明确，它希望取消拜登政府实施的许多人工智能监管，称他们想放开人工智能的束缚，赢得人工智能战争，领导世界。
At the recent AI Summit in New York attendees expressed excitement for the future of AI, but also a desire for policies and standards that will help mitigate its harms.
在最近于纽约举行的人工智能峰会上，与会者对人工智能的未来表示兴奋，而且还渴望制定有助于减轻其危害的政策和标准。
Companies want to understand how to avoid the kinds of footfalls and PR disasters that happen when, for example, AI is biased or AI fails to operate as designed.
公司希望了解如何避免发生的各种客流量和公关灾难，例如，人工智能存在偏见或人工智能未能按设计运行。
Gaming entrepreneur Jaye Watts notes the calls for transparency from both creators and companies like Valve, a major video game developer and distributor.
游戏企业家Jaye Watts指出，创作者和Valve等公司都呼吁透明度，Valve是一家主要的视频游戏开发商和发行商。
Now it is forcing the companies to say how they're using AI, instead of just using all these generative assets that you're getting from somewhere potentially infringing on someone else's right.
现在它迫使这些公司说明他们如何使用人工智能，而不是仅仅使用您从某个地方获得的所有这些潜在侵犯他人权利的生成性资产。
Ensuring that generative AI models are trained on quality data is ideal, but developers don't always have access to it.
确保生成式人工智能模型在优质数据上训练是理想的选择，但开发人员并不总是能够访问它。
When you use a pre- trained version of a large language mode, you even find tune, the last layer.
当您使用预先训练的大型语言模式版本时，您甚至会找到最后一层。
You do not have control on the massive amount of data it is already trained on.
您无法控制它已经训练的大量数据。
So there could be tuning but there's no concept of unlearning that data.
因此，可以进行调整，但别忘记这些数据的概念。
Experts suggest incorporating measurements and benchmarks to assess an AI program's performance, which they hope will ultimately drive more business.
专家建议结合测量和基准来评估人工智能程序的性能，他们希望这最终能够推动更多业务。
That's how you get buy in that's how you build in confidence.
这就是你获得信任的方式，这就是你建立信心的方式。
Others contend that a regulatory approach that's too heavy-handed could stifle innovation and competition.
其他人则认为，过于严厉的监管方法可能会抑制创新和竞争。
If some countries are more aggressive on regulations that some countries are less, then probably the tech companies will move across different countries.
如果一些国家的监管力度更大，而一些国家的监管力度较小，那么科技公司可能会跨越不同的国家。
AI, the field that every country want to win, so they will align their regulations.
人工智能是每个国家都想赢得的领域，因此他们将调整法规。
As artificial intelligence advances, human intelligence must still determine the right balance between innovation and order.
随着人工智能的进步，人类智能仍然必须决定创新与秩序之间的正确平衡。
Tina Trinh VOA News New York.
VOA记者蒂娜·郑纽约报道。
来源：英语东
