AI在行动：公用事业公司如何现代化电网运营

本季的日立能源PowerPulse播客，我们将聚焦电网自动化。

This season on Hitachi Energy's PowerPulse Podcast, we're turning our focus to grid automation.

这里有大胆的想法、真实的故事，以及来自正在塑造当今电网的专家们的真知灼见。

Expect bold ideas, real stories, and insights straight from the experts shaping today's power grids.

从智能数字解决方案到更可持续能源系统的挑战与机遇，我们将探讨最受关注的话题。

From smart digital solutions to the challenges and opportunities of a more sustainable energy system, we'll be covering the topics that matter most.

本期PowerPulse深入探讨AI如何从运营和智能层面改变电网。

This episode of PowerPulse dives into how AI is transforming the power grid operationally and intelligently.

随着AI驱动的增长加速电力需求，其影响有可能重塑我们所知的电网。

As AI driven growth accelerates electricity demand, its impact has potential to reshape the grid as we know it.

本期节目探讨AI如何塑造日立能源产品组合中的关键解决方案，以支持这一转变。

This session explores how AI is shaping critical solutions across the Hitachi Energy portfolio to support this shift.

我们邀请到了日立能源的数字战略负责人、AI领域的Milli Comstock，她负责领导AI和生成式AI项目，以现代化和优化电网运营，并帮助客户推进他们的AI战略。

We're joined by Milli Comstock, Digital Strategy Leader, AI at Hitachi Energy, who leads AI and Gen AI initiatives to modernize and optimize grid operations and help customers advance their AI strategies.

作为资深产品战略家，Milli以化繁为简、明确产品愿景而闻名，并致力于提供AI驱动的解决方案，以提升效率、可靠性和决策能力。

A seasoned product strategist, Milli is known for cutting through complexity to define clear product visions and for delivering AI powered solutions that enhance efficiency, reliability and decision making.

在本期节目中，我们探讨AI和数字创新如何在工业和工业物联网领域带来实实在在的影响，以及这些解决方案如何应对当前的挑战，同时塑造一个更可持续的未来。

In this episode, we discuss how AI and digital innovation are driving real impacts across industry and industrial IoT, and how these solutions are addressing today's challenges while shaping a more sustainable future.

和往常一样，我们会让对话保持清晰、实用，让所有人都能听懂。

As always, we'll keep the conversation clear, practical and accessible to all.

请享受。

Enjoy.

你好，Milli。

Hello, Milli.

欢迎。

Welcome.

嗨。

Hi.

你好吗？

How are you?

非常好。

Very well.

谢谢。

Thank you.

非常感谢你今天能来。

Thank you so much for coming in today.

我很高兴你能来。

I'm delighted to have you here.

谢谢你的邀请。

Thanks for having me here.

那么，Milli，我听说你是个很喜欢冒险的人。

So, Milli, I hear you're quite an adventurous person.

你有没有什么特别的冒险经历可以和我以及我们的听众分享？

Have you got a particular adventure that you could share with me and our audience?

是的。

Yeah.

我很乐意。

I'd love to.

我特别喜欢旅行，想跟你分享一个2007年的故事。

I love traveling, and I wanna share with you one of the stories back from 2007.

刚毕业那会儿，我和我丈夫想去巴塔哥尼亚旅行。

Right after graduation, me and my husband, we wanted to take a trip to Patagonia.

我们先到了布宜诺斯艾利斯，然后从那儿飞到了卡拉法特，那是通往巴塔哥尼亚的门户。

So we came to Buenos Aires and from there took a flight to El Calafate, which is the gateway to Patagonia.

我特别想看佩里托莫雷诺冰川。

My motivation was to really see the Perito Moreno Glaciers.

哇，真棒。

Oh, wow.

那冰川特别壮观，我们玩得很开心。

And it was incredible, and we had a great time.

但真正的冒险其实是那次临时起意的骑马活动。

But the real adventure was really this side impromptu excursion we took, which was the horseback riding.

说实话，我骑马完全是新手，我们报了个团，但导游不会说英语。

And I have to tell you, I'm a very beginner rider, and we bought a tour and the tour guides didn't speak English.

哦，哇。

Oh, wow.

而且我们不会说西班牙语。

And we didn't speak Spanish.

我们以为只会骑几个小时。

And we thought it was gonna be couple of hours.

然后旅程开始了，我们骑上马，然后，又有两个男的，对。

And when the tour started, we got on our horses, and there are, again, two guys Yeah.

他们很快教我们怎么让马往右或往左走，怎么拉缰绳让它停下来，这挺好的。

And give us this, you know, really quick how to tell the horse to go right or left, to pull the rein in, to make it stop, which is great.

但才过了五分钟，就有大概六七个男的和马加入了我们，对。

But five minutes into that, we were joined by, like, six or seven guys and horses Yeah.

他们还带着猎狗。

And they came with their hunting dogs.

我们当时就想，哦，这是怎么回事？

And we was like, oh, what was going on?

所以到头来，我们好像是报名参加了一次打猎活动。

So ultimately, looks like we signed up for a hunting expedition.

所以你可以想象——什么？

So you can imagine What

你们在猎什么？

were you hunting for?

你知道吗？

Do you know?

你们有没有——动物。

Did you did you ever Animals.

找到了，对。

Find Yeah.

我觉得我们确实有收获。

I think we did kills.

我没有，但队伍里确实猎杀了几只兔子。

I didn't, but the team did kill some rabbits.

但真正让人兴奋的是深入那片风景，是的。

But what was really exciting is going deep into the landscape and Yeah.

那是在智能手机或翻盖手机出现之前。

This was before smartphones or flip phones.

那持续了多久？

And how long was it

最后呢？

in the end?

差不多是一整天的活动。

Was a couple all day event.

等我们结束的时候，我浑身酸痛。

And by the time we were done, I was very sore.

可以想象。

I can imagine.

然后我们被邀请去吃晚饭。

And we were invited for supper.

当然，我们去吃了晚饭，那时候我还是个素食主义者。

And, of course, we went to the supper and I was a vegetarian back then.

所以呢

So

他们吃了兔子吗？

Did they eat the rabbits?

是的。

Yeah.

他们确实吃了。

They did.

好的。

Okay.

嗯，那听起来确实像一场冒险。

Well, that it does sound like an adventure.

你说得对。

You're right.

我很高兴我问了这个问题。

I'm glad I asked this question.

现在回想起来，说实话，我觉得挺不舒服的，我不希望我的孩子们去做那种事，因为我父母当时根本不知道我在那个地方。

Just looking back, you know, in hindsight, I mean, I find it very uncomfortable right now that it's not something I want my kids to go do because my parents didn't even know I was in that part of the country.

所以那真是一段难忘的记忆。

So it's very unforgettable memory

这倒是个好故事。

for Makes for a great story.

好故事。

Great story.

好，我们直接进入AI的话题吧。

Right, let's dive straight into AI.

你能告诉我为什么AI很重要，以及它在日立能源产品组合中的重要性吗？

Can you tell me why AI matters and why it matters in the context of the Hitachi Energy portfolio?

当然可以。

Absolutely.

AI是我们客户的一个战略重点。

AI is one of the strategic priorities for our customers.

我们有着悠久的运营技术传统。

We have this long heritage of operational technology.

我们是领域专家。

We are the domain experts.

现在你可以看到从消费级企业到日常生活的转变，人们已经习惯在日常生活中使用AI，无论是约会应用还是像ChatGPT这样的智能工具。

And now you see a shift from consumer enterprise where people, the workforce who are used to using AI as part of their daily lives, whether it's the dating apps or having access to smart tools like ChatGPT.

所以，我觉得人们有这种期待，为什么我们不能用AI来提升我们的日常运营流程呢？

So, I think there is this expectation that why can't we have AI to enable our operational workflow in our day to day lives?

然后我们看到了巨大的机会，通过行业里的一些基准测试，能够带来更高的效率和生产力。

And then we've seen a tremendous opportunity in terms of bringing more efficiency, increasing productivity from some of the benchmarks that have been done in the industry.

作为日立能源，我们听到了客户的声音，也做了一些早期试点，看看如何增强现有的产品组合。

So as Hitachi Energy, we have heard from our customers, we have done some early pilots in terms of how we can augment the existing product portfolio.

所以我们继续帮助客户实现他们的目标，同时在与同行和竞争对手的竞争中保持优势。

So we continue to help customers achieve their outcomes and still have that competitive edge against our peers and competitors in the market.

我们的主要客户之一是公用事业公司。

One of our main customers are utilities.

关于AI和公用事业的讨论主要在哪里进行？

Where is the conversation around AI and utilities happening?

那里的关键参与者是谁？

Who are the key players there?

讨论在几个地方进行，而且变得更加协作，也更加细致入微。

The conversation is happening in a few places and it's getting much more collaborative and much more reasonably nuanced as well.

例如，在美国，讨论主要由像EPRI这样的行业智库推动。

For example, in US, the conversation is much more driven by the industry think tanks like EPRI.

EPRI是电力研究协会。

EPRI is Electric Power Research Institute.

随着开放电力AI联盟的成立，它真正把公用事业、技术提供商、学术界和研究团体聚集在一起，来深入理解和探索AI在电力系统中的实际应用。

With the launch of Open Power AI Consortium, which is really bringing in your utilities, your technology providers, and academia and research groups to really understand and explore the practical applications of AI in the power systems.

日立是这个联盟的参与者。

And Hitachi is a participant of that consortium.

对吧？

Right?

而且你会看到很多公用事业公司之间的同行交流，他们分享一些试点项目和早期经验教训。

And you see a lot of sort of a peer to peer exchanges between the utilities where they are sharing some of the pilots and early lessons learned.

这些在美国都在发生，重点更多是电网的可靠性和韧性。

So that's all happening in US, and the focus is more on the grid reliability and grid resilience.

而如果你看欧洲，重点更多是围绕法规和脱碳目标，更强调治理和AI的可解释性，并且真正遵循欧盟AI法案的指导方针。

Whereas if you look at Europe, the focus is much more around regulations decarbonization goals, where the emphasis is much more on the governance and AI explainability and really following the guidelines of the EU AI Act.

当你看亚太地区，尤其是日本和韩国这样的国家，重点和势头更多是在自动化和先进的电网运营上。

When you look at APAC, especially in countries like Japan and South Korea, the focus and the momentum is much more stronger on automation and advanced grid operations.

这主要是由密集的网络和对更高可靠性的期望驱动的。

And that's primarily driven by the dense network and the expectation for higher reliability.

对吧？

Right?

在所有地区，对话已经从“AI能解决我们的问题吗”转变为“我们如何负责任地扩大AI部署，让人们能够信任它”。

Across all the regions, the conversation has really shifted from can AI solve our problem to how can we responsibly scale AI deployment enough so people can trust it.

所以这就是今天大多数有意义的工作发生的地方。

So that's where most of the meaningful work is happening today.

公用事业公司本质上相当保守，而且这是有充分理由的。

Utilities are quite conservative by nature and for good reason.

因此，他们以采用新技术缓慢而闻名。

So they're also known to be slow at adopting new technology.

你认为人工智能正在稍微改变这一点吗？

Do you think AI is changing that slightly?

是的，绝对如此。

Yes, absolutely.

而且可以理解的是，公用事业公司的首要任务是确保电力供应不断。

And understandably, utilities' main priority is to keep the lights on.

对。

Mhmm.

确保安全可靠。

Make sure it's safe and reliable.

我们现在看到的转变，其实是年轻一代进入职场，他们早就用上了消费级科技，不管是手机上的约会软件，还是像ChatGPT这样的工具。

And the shift we're seeing is actually the younger generation coming into the workforce who already have access to consumer technology, whether it's their cell phone using the dating apps or having access to tools like ChatGPT.

所以他们回到工作中就会问，为什么我们不能用AI来让日常工作流程现代化呢？

So they are going back to the work and asking why can't we bring AI to modernize our daily workflow.

对吧？

Right?

这种期望正在推动公用事业公司重新思考，AI如何增强他们的日常工作，带来一些能力，不只是为了技术本身，而是为了让工作流程更高效、更有效，减少人工操作。

So this kind of expectation is pushing utilities to rethink how AI can augment and bring some of those capabilities as part of their daily workforce, not just for the sake of technology, but to make sure the workflow becomes much more effective, more efficient, and less manual intensive.

另一方面，我们的专家也在陆续退休。

And on the flip side, what's also happening is our experts are retiring.

想想看，这些专家在十四、十五年的职业生涯中积累了多少机构知识，对吧？

So think of all the institutional knowledge our experts have accrued over the course of fourteen, fifteen years, right?

在这里，AI真的能发挥作用，帮助挖掘这些专业知识，进行知识转移，让年轻一代能够接触和利用这些知识。

Here, AI becomes a real enabler in terms of tapping into that expertise, doing the knowledge transfer and making it accessible and available for the younger generation.

所以这真的是双赢的局面。

So it's really a win win on both front.

那么，米莉，你能给我举几个具体的例子，说明AI是如何影响这些公用事业工人和终端用户的吗？

So Milli, could you give me specific examples of how AI impacts these utility workers and end users?

当然可以，我可以给你举几个AI发挥作用的例子，具体取决于你在整个能源价值链中的位置。

Yes, I can definitely give you a few examples where AI is relevant, depending upon where you are in the entire energy value chain.

我们以资产经理为例，他负责变电站的整体运营和维护，包括变压器、开关设备、断路器等等。

Let's take an example of an asset manager who's responsible for the overall operation and maintenance of a substation, which includes anywhere from your transformer, your switch gears, circuit breakers.

对吧？

Right?

有了AI驱动的预测工具，资产经理可以提前收到设备可能故障的预警，从而减少意外中断和停机时间，对终端用户的影响就是，你家的灯一直亮着。

So with the AI powered predictive tools, the asset manager can get an early warning sign of a potential failure to an equipment that reduces a disruption, any downtime that is not anticipated, and the impact to the overall users is you have the lights on all the time.

另一个好例子是，当你开车在高速公路上回家时，你左右看看，会看到高压输电塔。

Another good example may be as you're driving home on your freeway, you look to the right and left, you'll see high voltage transmission towers.

传统上，这是一个庞大的工程，需要训练有素的工程师开着卡车去现场。

Traditionally, it's a huge production where the trained engineers go with the truck roll.

他们爬上输电塔，手动检查，看看有哪些可能导致故障的风险。

They climb up the transmission tower and manually probe to see what are risks to failure conditions.

现在想象一下，无人机在输电塔周围飞行，从不同角度拍摄图像，这些图像通过预先训练的机器学习模型处理，再次给出故障风险。

Now take that with the scenario where drones are being flown around the transmission towers, and these drones are capturing images from different angles, and these images are being processed through pre trained machine learning models, which gives you, again, the risk to failure conditions.

但这样一来，你避免了很多责任问题。

But here, you have avoided lot of liability issues.

你提高了工人的安全性，而且所有这些都可以提前完成。

You have improved the safety of your workers, and all of these can be done ahead of time.

一旦你知道故障风险在哪里，你就可以部署工单，这些现场工作人员会带着平板或智能手机上的引导式移动应用去现场修复这些问题。

Once you know where are the risks to failure conditions, you can deploy work orders, and these field crews will go in the field and have this sort of a guided mobile app in their tablet or in their smartphones and fix those issues.

所以AI让你能够做很多这种主动的预先规划和调度。

So AI allows you to do a lot of these proactive pre planning and scheduling.

这反过来又影响了组织，因为降低了运维成本，因为你部署了正确的资源，确保正确的零件可用，这让一切更安全、更可用，而且周转更快。

This in turn impacts an organization by having reduction in your o and m costs because you're deploying the right resources, you're making sure the right parts are available, and that makes everything much more safer, much more available, and the turnaround is much more quicker.

你手头有没有一些实际的人工智能案例研究，以及你能分享的经验教训？

Have you got a couple of real world AI case studies in your portfolio and any lessons learned that you could share?

当然，我可以从我们的产品组合中给你举两个例子。

Sure, I can give you two examples from our product portfolio.

一个是针对我们现有产品APM的资产工作管理。

One on the asset work management for our existing product called APM.

APM是资产绩效管理，主要被我们的客户——公用事业公司——用来了解设备的资产健康状况，比如变压器、开关设备、断路器等等。

APM is asset performance management and is primarily used by our customers' utilities to understand the asset health of the equipment, whether it's a transformer, switchgear, circuit breakers, and so on.

但我们也知道，APM里数据量非常大。

But we also know that APM is inundated with lots of data.

有运营数据。

You have your operational data.

有传感器数据。

You have your sensor data.

还有可靠性模型，会生成各种资产健康风险评分。

We have reliability models that is spinning out all sorts of asset health risk scores.

对吧？

Right?

所以对于APM用户来说，有时候要找到做决策所需的关键信息，会花很长时间。

So for APM users, sometimes trying to get to the critical information required for decision making takes a lot longer.

实际上你得翻好几个界面，还得建仪表盘，提取数据，再用Power BI做关联分析，这些麻烦事都得做。

You will actually have to go through multiple screens, you have to create dashboards, you have to extract data and then correlate using Power BI and all that good stuff.

所以我们想用生成式AI的能力，帮大家更快地获取这些信息。

So we wanted to use Gen AI capabilities to help get to those information faster.

这次我们和硅谷一家叫Bolo AI的初创公司合作。

And in this case, we are partnering with a startup based in Silicon Valley called Bolo AI.

和Bolo AI合作，我们用的是ChatGPT 4.0，并在上面加了一个语义层，让它能更好地理解APM数据。

With Bolo AI, what we're doing is we are using ChatGPT 4.0 and providing a semantic layer on top so that it understands understands the APM data better.

它明白资产健康评分到底意味着什么。

It understands what does the asset health score really mean.

对用户来说，好处就是你不需要知道APM是什么。

The value to a user is you don't really need to know what is an APM.

你也不用去建各种表、做各种关联查询来获取数据。

You don't have to go create different tables and joins to get to the data.

你只需要问一个简单的问题，比如：哪些是风险评分最高、故障概率超过80%的前五大关键变压器？

You can ask a simple question like, what are some of the top five critical transformers showing highest risk scores and the probability of failure is above eighty percent?

像这样的问题，不到两秒钟就能得到数据点。

For something like that, it just takes less than two seconds to get to the data point.

这可以成为对话的开端。

That can be the start of the conversation.

你可以继续向模型提问，模型会继续深入挖掘，最终给你一个结果。

You can continue to ask the model and the model will continue digging deeper and get you to an output.

我们还与一个叫Emron的客户合作，Emron的工程师们会问一些问题，比如“给我看一下SF6气体的报告”。

We also work with a customer called Emron, and some of the questions that Emron engineers are asking are, show me the report of SF6 gas.

对于一位训练有素的Emron工程师来说，制作这份报告每个月要花八到十个小时。

And this takes anywhere from eight to ten hours per month for a trained expert Emron engineer to produce this report.

这又是因为你需要做大量的数据导出，在Excel里做数据透视表，然后发给领导层。

Again, because you are doing a lot of data dump, doing pivot tables in Excel and then circulating amongst your leadership.

Emory提出的另一个很好的用例是乙炔峰值检测，这是变压器故障的早期信号。

Another good use case that Emory brought up was really on the acetylene spike detection, which is an early sign of transformer failure.

这些都是APM用户真正想问的问题，他们不需要深入了解APM，也不需要花大量时间在APM里翻找信息。

So these are all the real world questions APM users want to be able to ask without really understanding and without spending a lot of time digging through the information within the APM.

那太好了。

That's great.

那为什么与初创公司的合作特别重要呢？

And why are partnerships with startups so particularly important?

这个问题问得非常好。

That's a really great question.

我觉得合作伙伴，尤其是初创公司，能带来那种灵活性、专注度和规模化能力。

And I think partners, especially the startups, bring that agility and focus and scale.

我们清楚客户的痛点在哪里。

We know what the customer's pain points are.

我们知道自己在运营技术方面是专家，但像我们合作的这些高度专业化的初创公司，比如一家叫 Articulate 的初创公司，他们在这方面非常擅长。

We know we are expert when it comes to operational technology, but there are highly specialized startups like the ones we partner with, in this case, a startup called Articulate.

这实际上是在解决我们网络控制产品组合中最大的挑战之一，也就是自动化单线图生成流程。

And that is really working with one of our greatest challenges in the network control product portfolio in terms of automating the single line diagram generation process.

目前，我们的网络控制团队在进行升级或新系统安装时，可能需要花一百多个小时来查看所有 PDF 中的示意图。

Today, our network control team, when they have to do upgrade or new system installations, it can take anywhere from hundred plus hours to look at all of the schematic of the PDF.

单线图可能是JPEG格式的。

The SLD could be in JPEG format.

可能是不同年代创建的。

It could be created over different eras.

Articulate真正帮到我们的地方，就是把这个过程自动化了，包括提取数据，以及和底层网络拓扑做数据映射。

How Articulate really helped us is automate the process for us in terms of extracting, doing the data mapping with the underlying network topology.

而额外的优势就在于，领域特定的大语言模型在这里发挥了作用。

And the added benefit is where domain specific large language model comes into play.

现在我们可以根据生成的新单线图，提出具体的问题，对吧？

We're able to now ask specific questions based on the output of the new single line diagram that is being produced, right?

这不仅仅是和Articulate的合作，也回到了我之前提到的Open Power AI联盟。

And this is also not just touching on the partnership with Articulate, but it also goes back to what I mentioned earlier as part of the Open Power AI Consortium.

Open Power AI联盟的目标之一，就是创建这种领域特定的模型，这和通用的大语言模型很不一样，因为这些模型是基于能源、电力和公用事业数据训练的。

One of the objectives of the Open Power AI Consortium is creating this domain specific model, which is very different from the general purpose large language models because these models are trained on energy and power and utilities.

所以准确度更高，而且对上下文的理解也更深入。

So you have higher level of accuracy and it's much more context aware.

Milli，那个

Milli, the

西南电力池和日立能源的公告突出了像并网排队、输电建设缓慢这样的瓶颈问题。

Southwest Power Pool Hitachi Energy announcements highlighted bottlenecks like interconnection queues, slow transmission build outs.

你们是怎么用AI来应对这些挑战的？

How are you addressing these challenges with AI?

在我之前的回答里，你知道，谈到与西南电力池合作的并网排队研究，大概需要十八到二十四个月，这个挑战非常大，不是单靠一家公司就能解决的。

On my earlier response, you know, talking about partnership with Southwest Power Pool's challenge of the interconnection queue study taking about anywhere from eighteen to twenty four months, the challenge before is huge, not just one company can address it alone.

对吧？

Right?

这就是为什么我们和NVIDIA合作，他们是一家很棒的数据科学公司。

That's why we are partnering with NVIDIA, which is a great data science company.

我们和SPP合作，他们是这个过程中的客户，而且不只是日立能源，还有其他姐妹公司，比如日立美国、日立数字、日立Venterra。

We're partnering with SPP who's a customer in the journey and not just Hitachi Energy, but other sister organizations like Hitachi America, Hitachi Digital, Hitachi Venterra.

我们正在非常全面地看待这个问题。

And we are looking at this problem very holistically.

我们正在研究如何解决这个问题，不只是增加一个AI功能，而是采取一个非常全面的方法，从NVIDIA GPU计算开始，利用日立美国、Hell R&D的AI替代模型，同时增强我们现有的ProMod，真正实现以前需要四十小时的模拟，也许能缩短到四小时。

We're looking at how we can solve the problem, not just by adding one AI capability, but taking a very holistic approach, starting with NVIDIA GPU computing, having Hitachi America, Hell R and D's AI surrogate models, and we are enhancing our existing ProMod to really perform the simulation which used to take forty hours, maybe cut down that by four hours.

所以，在这些解决方案中嵌入了不同层次和级别的优化和AI，为客户提供全面的结果，我们希望将这一巨大挑战减少大约80%。

So there's a different tiered and level of optimizations and AI embedded into these solutioning to give that holistic outcome to the customers, which we hope to cut down this huge challenge by about 80%.

我们还在努力中，也许几个月后，我们可以带着这个成果回来。

We're still in the works and maybe in a couple of months, we can come back with that outcome.

那么，Milli，展望未来，比如说三到五年，你认为AI会把公用事业带向何方？

So Milli, looking ahead, let's say three to five years, where do you see AI taking utilities?

在未来三到五年内，你会看到很多AI被部署在控制室操作中。

In the next three to five years, you will see a lot of AI deployed in the control room operations.

今天，很多决策是由操作员根据大量历史数据做出的。

Today, a lot of the decision making is done by an operator who's looking at a lot of historical data.

但是有了AI，你会有AI代理，会有更多实时数据涌入。

But with AI, you will have AI agents, you will have more real time data coming in.

AI将能够根据预期的中断实时调整。

AI will be able to pivot according to anticipated disruptions in real time.

所以最可能的情况是，AI代理会四处运行，做出一些运营决策，但人类仍然会参与其中。

So the scenario would be most likely the AI agents running around making some of these operational decisions, but human will still be part of the process.

人类操作员的角色会提升到更多的监督层面，确保所有这些AI代理在做关键决策时是安全可靠的。

The human operators rule be elevated to more of a supervisory role to ensure that all of these AI agents are making the right decisions, critical decisions in a safe manner.

对。

Mhmm.

所以本质上，你会看到一个更加自主和自愈的电网。

So in essence, you will see a grid that is much more autonomous and self healing.

这样就能继续为所有人提供可靠和可持续的能源。

So it continue to provide reliable and sustainable energy for all.

我最后一个问题。

My last question to you.

你希望你的客户或团队在AI方面多问些什么？

What is the one thing that you wish your customers or your teams were asking about AI?

也许我可以换个角度来回答这个问题。

Maybe I can take that question a little bit differently.

是的，很多客户和同行都在问，我们怎么用AI？

Yes, a lot of customers and my peers are asking, How can we use AI?

我觉得问题应该是，你想解决什么问题？

I think the question should be, What is the problem that you're trying to solve?

对我来说，AI只是达到目的的一种手段。

I think to me, AI is a means to an end.

AI是实现某个结果的推动力。

AI is an enabler to an outcome.

所以我们应该先聚焦于问题本身，搞清楚问题是什么，再考虑解决方案。

So we should be focusing on what the problem is, getting down to that before jumping into the solution.

当然，AI和技术可以解决任何问题，但我们要确保先定义好问题，同时也要明确期望的结果是什么，而AI能帮助我们达到那个目标。

And of course, AI and technology can solve any problem, but we want to make sure that we are defining the problem, and we're also defining what the outcome should look like, and AI can help us get there.

和你聊天真的很愉快。

It's been wonderful chatting with you.

非常感谢你今天抽空过来。

Thank you so much for making the time to come in today.

谢谢你的邀请。

Thanks for having me.

今天的内容就到这里。

And that's it for today.

我们很快会带来更多精彩内容。

We'll be back soon with some more great content.

在离开之前，记得关注我们，这样就不会错过任何一期节目。

But before you go, remember to give us a follow so you don't miss an episode.

感谢收听。

Thanks for tuning in.

我们下次见。

See you soon.

本期节目由日立能源赞助播出。

This episode was brought to you by Hitachi Energy.

由芭芭拉·弗雷塔斯·丹尼尔斯创作、介绍并主持。

Created, introduced, and hosted by Barbara Freitas Daniels.

内容与脚本撰写由米莉·科姆斯托克负责。

Content and script writing by Milli Comstock.

制作与剪辑由创意猩猩团队完成。

Produced and edited by Creative Chimps.

本期嘉宾是米莉·科姆斯托克。

Guest speaker, Milli Comstock.