チャットGPT、生成AIの登場で経営戦略はどう変わる？～安宅和人×松尾豊×上野山勝也

[Click here for a summary of this video] https://mba.globis.ac.jp/knowledge/de... G1 Venture 2023 Part 8 Plenary Session: "Globally Competitive Technovate Management: A New Management Strategy Methodology Utilizing ChatGPT and Other Technologies" (June 11, 2023 / Globis University Graduate School of Management, Tokyo) Kazuto Ataka, Yutaka Matsuo, and Katsuya Uenoyama This year's G1 Venture focused on the theme of "Globally Competitive Technovate Management: Utilizing Generative AI and Other Technologies," delving into the evolution of technology and management from every angle. How will generative AI, such as ChatGPT, change the way businesses manage their businesses, and how can companies leverage new technologies to gain a competitive edge? The final plenary session explored new management strategies in the age of AI. (Titles are as of June 11, 2023) -------------------------------------------------------------------------------- ✅To learn more about ChatGPT, click here. "Start Using ChatGPT Tomorrow! Introduction to Using ChatGPT" https://hodai.globis.co.jp/courses/aa... ---------------------------------------------------------------- ●Kazuto Ataka (Professor, Faculty of Environment and Information Studies, Keio University / Senior Strategist, Z Holdings Corporation) ●Yutaka Matsuo (Graduate School of Engineering, The University of Tokyo) Professor) ● Katsuya Uenoyama (CEO, PKSHA Technology, Inc.) 00:00 Opening 00:21 Thoughts since the discussion at the G1 Summit three months ago (Related video:    • チャットGPTは何がすごい？今後、何が起きる？～安宅和人×松尾豊×平野未来×上野山勝...  ) -A lot has happened in just two months: the Ministry of Education, Culture, Sports, Science and Technology guidelines, the University of Tokyo statement, the AI ​​Strategy Council's establishment and discussions, and the Hiroshima AI Process at the G7. -We've entered an era where everyone knows about LLM. ChatGPT has the same usage time as Twitter, so you could say it's the emergence of a giant media outlet. Single-source products are emerging at an extraordinary speed. We are truly living in a historic moment. Apple Vision Pro has also been released. 03:46 How should management perceive the current environment? Where will change begin and how? -Text generation and brainstorming are becoming easier. RPA and other internal processes can be significantly automated. -How many layers will LLM have in the future? Will it be one? Or multiple? -Things like hallucination will barely be a concern in a year. The advent of traditional machine learning and deep learning has almost completely eliminated language barriers. Just when diffusion models were emerging and images and music were becoming available, LLMs are now connecting with robotics. It's impossible to understand without thinking about this 4+1. Focusing solely on LLMs will lead to failure. Speaking of industry, there's the old economy, centered on hardware; the new economy, which lives in the data world; and a third force like Tesla and Uber, which are using cyberpower to solidify the real world. It's dangerous to discuss these things together. In any case, LLMs will certainly assist white-collar work. However, the real problem is that many people are arguing that their jobs will disappear. The discussion should instead focus on how to create a new future. -Most value-added work is like a PDCA cycle. However, there are companies that actively use generative AI and others that consider it dangerous and put the brakes on it. What's holding them back? -Perhaps it's a mentality. Students at Matsuo's lab bring new ideas to the table, and Mitoyo City in Kagawa Prefecture has released a GPT trash collection guide. 12:56 Perhaps human behavior is already being altered by algorithms in "employment practices" and "personnel systems." Japan is a country where air weighs more heavily than law. The biggest burden is not following through on what can be done by law. It's impossible to plan a budget when things move so quickly in a single year. Both universities and the government operate on a one-year budget, which is completely out of step with this. The root cause is a society that doesn't control things by results. The idea that checking off a checklist is enough is no good. Yahoo!'s Ozawa has said that the presidents of listed companies will become AIs, and in Silicon Valley, a company has already pitched ChatGPT and completed fundraising. If AI were to handle personnel systems and one-on-one meetings, perhaps interactions could emerge. 18:44 From a technical perspective, what are the limitations of GPT and how will it change in the future? How far will the current scale of GPT go? This will require more data, and it remains to be seen how accurate it will be. Some kind of technological breakthrough will be necessary. -Regarding the risks of manipulating the real world. I think people are worrying too much about it bec...