Artificial intelligence is coming, what changes need to be made in the way of work and management?

Original source: 36 krypton

Image source: Generated by Unbounded AI‌

Editor's note: Microsoft is working closely with OpenAI to bring generative artificial intelligence capabilities to many of the tools used in the workplace, including new features like Microsoft 365 Copilot. Its goal is to change the traditional way of knowledge work. In the coming years, artificial intelligence will have a transformative impact on work and management. For example, artificial intelligence can help improve communication efficiency, find focus, improve creativity, and meetings will become queryable knowledge objects, etc. Artificial intelligence will change management philosophy, organizational structure and role setting, and managers need to adapt to the development of new technologies. Most importantly, we must maintain a correct attitude towards the emergence of new technologies and cultivate the ability to adapt to new technologies. The article is from compilation.

In the past few years, Microsoft has published some research on how people work, including an analysis of the surge in meetings since the outbreak and the phenomenon of the "three-peak" workday, which is very interesting.

Now, based on a close partnership with OpenAI, Microsoft is adding new capabilities to many of the tools used in the workplace, fully introducing generative artificial intelligence capabilities. These new capabilities, which include Microsoft 365 Copilot, are designed to further change the way knowledge work is done.

To better understand how this affects work and management, we caught up with Jared Spataro, vice president of modern work and business applications at Microsoft. The following are excerpts from our conversation, edited for length and clarity:

**How would you summarize where we are in the evolution of knowledge work? How do you see our jobs going to be in the next two to three years as AI tools roll out? **

We are in the early stages. In many ways, the past few decades have been the digitization of analog working models. For example, a spreadsheet is a digital representation of a ledger. Word processing evolved from the typewriter. You'll see that we've digitized the physicality of these analog states. In many ways, the benefits of this approach have diminished marginally, and now we are beginning to enter a new era of true digitalization. We're really starting to think about what we can do differently when we're actually going digital. We realized, especially because of the pandemic and the massive changes in all different types of communication, that we could apply AI to a lot of tasks that we thought about and thought, well, this is the way to go. So that's very exciting to me because we're coming out of the first phase of digitization and really entering a new, very exciting phase.

**In our work, what do you think we are doing now that we will stop doing in a few years? **

If we look back and look at what AI has done for me and people who are just getting started, we see some patterns. Number one, we do see AI helping us find the signal in the noise. We found in the actual statistics that a very interesting fact is that the average employee spends almost 60% of his time communicating or coordinating with others every day. And this proportion is still growing. Everyone can definitely feel this. The way they express it is that I have a lot of work to do before I can do my real job, which is what they've hired me to do. So, I see a pattern emerging where AI is going to help us solve all of these problems.

Not that there will be less communication. But we'll be able to figure out where the communication is most important to help us get the job done. I am also very interested in how artificial intelligence can help us stimulate creativity. We will see the emergence of entirely new models of human-machine collaboration. I don't think it means we will collaborate less with humans, but AI will increase our creativity. For me, this idea is unique. For years, I've believed that machines are bad at creative tasks. In the end, conferences can change a lot. It's not that we don't need to have meetings anymore, but I've found that, for example, I'm attending fewer and fewer meetings because then the meeting is no longer a point in time, but an object of knowledge that I can query and ask questions . It's a whole new way of thinking about human interaction. So those three things are going to change a lot in the next few years.

**Artificial intelligence enhances employee creativity? Can you expand on this in detail? **

On the surface, artificial intelligence can flesh out your rough ideas. Sometimes it's really a daunting task for everyone. They have a spark of thought, and they want that spark of inspiration to become something flesh and blood so that they can imagine the idea in a more complete form. I do see this starting to happen now. What I appreciate most, though, is that there can be a real spark that can even be traced back to the creative process. I've found that over the years I've been using services, agencies, and people to help me generate ideas. It's almost like a funnel of ideas, and I'll gradually narrow it down until I lock in on what I want to pursue with the work. I've found that it's actually possible to widen the top of the funnel. I can generate ideas faster. It also really fundamentally changed the way I thought about the early creative process. Again, the most surprising thing to me is that I didn't even think about this a few years ago. Honestly, I think AI is better at automating mundane tasks than generating ideas.

**You speak of meetings as "knowledge objects". Is it because AI can easily transcribe and summarize? **

Yes, but there is another side. Transcription technology is getting amazingly good. It's not that there are no errors at all, but even if people do the transcription, it is not without errors. In my experience, artificial intelligence has far exceeded human capabilities. The session wrap-up is very valuable, but for me it's the next step. It's good to have a summary. A summary can give you an overview of things, but you usually want to know more about human interactions. For example, I'm amazed that in some of the tools we're releasing now, you can ask very specific questions. Has anyone mentioned my name? Has this topic been mentioned before? What does the team think about this? What decision did they make? How do they feel about these decisions? Are there any disagreements? You can analyze interpersonal interactions, and this level of detail will open your eyes. Wow, in the past, you would just jot it down after the meeting, but now, you can look at this very comprehensive information from many different angles. interactive.

**Both real-time and retrospective? Does the current technology have the conditions? **

It does, and both are excellent. Meeting retrospectives were one of those things where, "Wow, I don't think I have to go to every meeting to make sure I'm on track." It changed my pattern of meeting attendance. But our team told me that you have to see it to believe what this technology can do for meetings. In fact, my product team and I had a very friendly debate about this. They said, “Jared, you gotta watch this. It’s going to change the way we conduct meetings, and it’s really worth watching.” I told them, “Guys, the real value of meetings is after the meeting.” And they said : "No, you try it first." When I tried it for the first time, I was really surprised, because the current GPT model is very good at doing this, and can even determine the emotion of the speaker based on real-time transcription. Every time I show this to a client, it's always very impressive.

**Going back to the previous topic, which functions or characteristics of artificial intelligence do you think will have the greatest impact on knowledge workers in the next few years? **

This question is not easy for me. When people ask me that question, it's like, "Oh my god, Jared, I bet you've got a whole bunch of them, but you're not even sure." No, I'm clear. The biggest impact is the emergent nature that tools like ChatGPT bring to your business. ChatGPT is amazing because if you ask it questions related to its training material, it will give reasonable answers, surprisingly reasonable, sometimes even wrong. We can talk about hallucinations, but the answers it gives are surprisingly plausible. These answers can help you if you're trying to do a specific task. Today, we don't have the equivalent of ChatGPT for your business that you can ask for everything from Q4 sales to current trends to forecasts for a particular product line at the end of the quarter.

There is no such thing yet. We have no competition in this regard. This technology is emerging now, and it will appear much faster than people think. It will reshape job design, roles and, dare I say, even how companies operate. Because a lot of the roles are essentially about delivering information and moving the company forward in order to make the right decisions about some of the key aspects of making a product and selling it. I think that's the way it is. I think people who do finance, marketing, sales will spend less time on the traditional applications that have emerged in the past few decades, and they will spend more time looking up the company's situation, hey, how is this situation? What about that? What if we did this? How should I think about this problem? Do you have any suggestions for this? This is a new way of thinking that focuses on the ongoing operation of a business.

**You said it's going to happen much faster, is that a year or two years from now? **

Much faster than that. We came up with these ideas in March of this year. You can expect to easily use it in production this fall. By the end of this summer, we'll be showing customers some of the things that can be done with it. The technology is there. In the final analysis, the basic idea is to use the GPT model to do the so-called "grounding" operation that we introduced a few months ago. That is to connect your data to the model. This is a very efficient way to integrate your data with the GPT model. This is the architecture of the co-pilot, which is being developed by everyone, and we are now co-developing our own architecture.

**Given that new AI tools can hallucinate and make mistakes, is it okay to allow a computer to make mistakes some of the time and let it do its thing? Do we need to make a decision on this? Will this be the case in the foreseeable future? Shouldn't we feel uncomfortable when this happens? **

Yes. The term "actually wrong" that we create and process when we make products and work with customers. In fact, we tell people when we're wrong, but the mistakes often have value. Just like the people you work with may have a less-than-perfect idea or a less-informed point of view, but those back-and-forths are what drive the value. What I'm trying to say is that I've made a very important discovery or insight over the past few months: Most people are used to using computers like calculators. I would ask some kind of definitive question. You give me the answer to this question. You should give me the correct answer every time. And with these new fundamental big language models, you basically have a general purpose inference engine. You give it information and give it hints or questions. It will reason about this information in the best possible way and try to find answers to your questions. Answers can also be wrong. What we're going to tell you right now is that even if the answer is wrong, it's very interesting to track down why it came about. How does it fit the pieces together? How might this advance your own thinking? This is completely different from asking arithmetic questions with a calculator.

**As jobs evolve over the next few years, what will managers need to do differently to be successful? **

Well, we're at a real inflection point in management. Of course, we've been chanting this for decades, but the management philosophies, the organizational structures, the theoretical perspectives—the various models we learned in business school—are essentially based on the organizational methods of the post-World War II industrial complex. Command and control is at the top of the list, and it's one of those very large practices. But beyond command and control, even some of the smaller, more nimble organizations largely mimic traditional structures. I think artificial intelligence will upend this status quo. Of course, distributed offices and flexible working have upended this. So I think when we look back in five, 10 years from now, we'll say, wow, the new managers are completely different people with all these technology confluences. They need to know how to manage the time and energy of the people in the organization across time and space. They must be able to recognize that augmenting human capabilities with machine-based capabilities will help them do their jobs faster, better, and with higher quality. All of these factors are things that we never really think about. So while I can't predict exactly what that will look like, it's easy to predict with 100% certainty that the managers of the next two or three years will look similar to the prototypes trained in business schools over the past few decades. there is a big difference.

**Companies are asking executives for their views on the impact of AI on talent strategies, especially in the context of 2024 budgets and strategic plans. How should they respond? **

The best model I can see right now is to jump in, start using it, form your own opinion, be willing to try the technology, and understand its capabilities and limitations. If you want to be more specific, my suggestion is that, for example, you should let the marketing department actively experiment with artificial intelligence, from targeting marketing (how to isolate your marketing goals in the data) to content generation (of course content can be generated very efficiently) and even creative tasks. If you're looking at the cross-organizational sales function, I'd have salespeople leverage AI to help them engage with customers more effectively and find the right information in the sales process. Essentially, every function we can take a closer look at, most of the visionary companies we see are currently looking at how AI can be applied function by function.

And then, instead of copying everything that's written in the media, they're going to look for the essentials of state-of-the-art technology, and they're going to say, "That's great. I think we can reduce some resources, let's say, to copywriting Reduce it by 10% and see what happens." Or "I think we can use generative AI on images to help us with graphic design. Let's embrace AI, not fight it." In my view Come to think of it, this is often the most reasonable strategy at the moment. No one knows exactly how the future will develop, so you have to have your own feelings and judgments.

**Do you expect artificial intelligence to bring job growth, both qualitatively and quantitatively? How can people and organizations ensure that this happens, rather than having AI automate and replace human jobs? **

There is a saying in economics called "the lump of labor fallacy" (the lump of labor fallacy), which means that the total amount of work that needs to be done in society is fixed. If you replace part of it, some workers will be replaced. Lose. The good news is that over the years it has been proven to be a fallacy. Nowhere in the world has this happened historically. When we invent labor-saving devices, the economy finds a way to use them, and when these devices are used, the economy grows with the devices, and as the economy grows, jobs are created. So I'm not concerned about what this technology will do to the labor market and our economy in the medium and long term. I feel great. In the short term, human labor will definitely be replaced. Therefore, it would be disingenuous to say that in some cases the effect of artificial displacement will not occur.

But from my perspective, when you look back at the history, the pattern is very clear. What should you do when a disruptive technology emerges? embrace it. You're sure this usually ushers in a very interesting era of growth drivers, and the best way to adapt to this era is to make sure you learn. I have to adapt to this era. Humans are very adaptable. So even in the short term, as an individual or as an organization, if you take this approach to artificial intelligence, you can do very well, because this is essentially a wave of force that is sweeping through the economy. So personally, as a company, we're very optimistic about technology in general. We really feel that you shouldn't do things with fear. You should of course look at it with respect and think about the limitations of this technology, but together we can do a lot if we embrace this technology.

**Historically, underrepresented groups, including women, have been left behind during technological change. What do you think needs to be done to make sure this doesn't happen again? **

I'm going back to the very powerful idea that when new technology comes out, you have a choice to decide how to use it. One of the exciting things about AI that we're seeing is its broad application. ChatGPT is the fastest growing consumer technology we've seen. Hit 100 million users in less than 90 days; today, there are over 200 million users. So when it comes to penetration, you'll see that's the widest penetration we've ever seen. Of course, we want to do more in terms of pervasiveness, which is why we are funding OpenAI so that anyone, anywhere in the world, can use it for free, for any use case they want to address. But at the same time, in addition to usability, you also have to teach people the right attitude. It’s a growth mindset, I would say. It’s a mindset of embracing and stretching. If you see an opportunity for technological change, you can take advantage of it.

**You mentioned the concept of "AI talent" in your latest study. How is it best to teach AI talents when AI itself is developing so rapidly? As a skill, how do you ensure its sharpness? **

This is one of the challenges facing technology today. I mentioned the meteoric rise of ChatGPT. We've never seen anything like it. The good news is, what I've found in my interactions with the emerging generation is that they embrace this stuff very quickly. I think that's what you have to do, teach people how to embrace new technologies. I'm heartened when I see educators telling students, let's embrace this technology, not fight it. what can you do? Let's learn how to use it to become a better writer or a better thinker. I encourage society as a whole to do the same. Don't run away from these techniques. Don't worry about kids using them to cheat. Instead, we decided we could collaborate together and learn how to use the technology effectively. Most importantly, don't do things out of fear. To carry out work with optimism and an attitude full of opportunities. If you do that, not only will things go well, but we'll also be able to take advantage of the surplus that these technologies generate, and they can be used on a global scale.