The forefront of the AI revolution: How will jobs, skills, and mindsets change? – Sushi Tech Tokyo 2025 "The dawn of the AI era"

Introduction: Exploring our future as AI advances at an accelerated pace

On May 8, 2025, Tokyo Big Sight was filled with excitement. On the main stage of "Sushi Tech Tokyo 2025," where future technologies come together, one session that particularly attracted attention was "The Dawn of the AI Era: How Will Our Work Change?" Three experts working at the forefront of AI took to the stage and engaged in a lively discussion about how the evolution of AI will bring about changes to the way we work and society.

This report provides a comprehensive explanation of this stimulating session, with over 5,000 characters based on the transcript and materials provided. From the latest trends in AI to changes in actual business situations and the mindset we should have, we will deliver insights that will serve as a compass for the AI era through the speakers' live voices. This will be a must-read for anyone who has hopes and some anxiety about the evolution of AI.

Speakers: Top runners in the AI industry

This session was highlighted by the following three speakers, all of whom are leaders in the social implementation of AI technology in their respective fields.

• Mr. Masahiro Chaen: CEO of Digirise Inc. An expert in corporate AI training and ChatGPT development for businesses. He has over 150,000 followers on X (formerly Twitter), and is a huge influencer in disseminating information about AI. He also serves as an advisor to several companies, including GMO AI & Web3 Inc., and is deeply familiar with the business use of AI.
• Aoi Kobayashi: CEO of Trickle, Inc. A young entrepreneur who heads Trickle, Inc., a company that develops and provides training on generative AI tools while studying at the School of Political Science and Economics at Waseda University. She is also the founder of the Waseda AI Research Group, one of the largest generative AI research groups in Japan, and is attracting attention as a female opinion leader in the field of generative AI and development.
• Maiko Kojima: Representative Director of Craftar Inc. He is involved in the planning and operation of "Crew," a secure AI tool that can reference proprietary internal documents, and "CraftChat," a digital transformation sales promotion support tool, and in 2022 achieved M&A with the Monex Group. He is active in both AI business and diversity promotion, including being a member of the General Incorporated Association for the Promotion of AI Utilization (GUGA) and being selected as one of the Forbes JAPAN Women In Tech 30.

Under the moderator's skillful guidance, the three panelists shared their knowledge and offered many hints for surviving the age of AI.

Part 1: AI Startup Trends and Frontiers – Where is the Technology Headed?

The session began with "Trends and the Frontline of AI Startups," which looked at the current state and future of AI technology. The speakers keenly grasped the wave of change that is currently occurring and spoke about its core.

1-1. Three keywords that will dominate the AI industry in 2025: inference models, agents, and creative AI

Chaen began by pointing out three major current trends in AI.

• Evolution of inference models: "Since around the end of last year, models such as those from OpenAI have appeared that have improved response speeds and accuracy not just in the development stage but when we actually use them. This has made them much more practical, and I feel like we've made a breakthrough," said Chaen. Models that were previously high-performance but heavy-duty can now be used for inference more efficiently, expanding the scope of their applications.
• The Rise of Agent AI: "As you all know, agent AI that automates browser operations is attracting attention. If this becomes widespread, it will automate many clerical tasks. In the second half of 2025, it may be standard on browsers such as Chrome, and we may be able to give instructions such as 'send an email to Mr./Ms. X' or 'compare this product and purchase it,'" (Chaen). This suggests that AI will evolve from a mere provider of information to an 'agent' that carries out our instructions.
• A breakthrough in creative AI: "The accuracy of image and video generation has improved dramatically. Figma announced its AI features today (the day of the session), and AI will likely make it easier to create websites and other things. By next year, we may be able to create YouTube thumbnails at a level where experts will no longer be needed," said Chaen. The evolution of AI in the creative field has the potential to fundamentally change the way content is created.

1-2. Evolutionary stages of AI agents: Where are we now?

Kobayashi provided a more detailed perspective on the evolution of agent AI in particular. She sees the evolution of AI agents in five stages:

1. Chat AI: AI that can have natural conversations with humans (e.g. ChatGPT).
2. Step-by-step thinking AI: An AI that can understand a user's problems and provide step-by-step solutions.
3. Tool-linked AI: AI that can link with external tools such as Google Calendar and email.
4. Autonomous research AI: The stage where AI can conduct research on its own.
5. Near-AGI (Artificial General Intelligence): An AI with intelligence equal to or greater than that of humans.

"We have now fully reached the third stage, 'tool-collaboration AI,' and in some areas we are moving into the fourth stage, 'autonomous research AI,'" said Kobayashi, emphasizing that AI is rapidly improving its capabilities.

1-3. Current status and challenges of AI use in Japanese companies

Meanwhile, Kojima also mentioned issues unique to Japanese companies, contrary to the global trend. "The reality is that Japanese companies often do not have their own company data online, as is the case in the US. They are still in on-premise environments or manually uploaded to file servers. In order for AI to demonstrate its true value, a major challenge is how to smoothly link this data with AI," said Kojima. The point that in order to maximize the capabilities of AI, it is essential to first improve the data environment at hand may be hard to hear for many Japanese companies.

1-4. AI tools used by professionals and how to use them: ChatGPT, Claude, Gemini, Gamma, Notion AI…

The discussion then moved on to more specific topics, such as what AI tools the speakers actually used and what results they were achieving.

Chaen says he skillfully uses multiple AI tools. "For document creation, which is common among Japanese companies these days, Gamma is the best. Its development speed is very fast, and its functions are constantly evolving. Of course, I also use ChatGPT and Claude, and I also pay full fees for Google's Gemini. Depending on the purpose, I use ChatGPT for file analysis, and Claude for long-text reading and internet searches," says Chaen.

While Kobayashi agreed with the usefulness of Gamma in the business world, he added his own perspective as a student: "I use PMDV and PEEK to transcribe meetings, but as a student, I find Notion AI extremely useful for taking notes in class. It transcribes in real time and automatically summarizes what I've said, so my learning efficiency has improved dramatically."

Mr. Kojima said that while the use of ChatGPT is spreading rapidly among university students, he has the impression that its use by working adults has yet to take off, and he also touched on the gap in AI usage between generations.

This segment demonstrated just how exciting the cutting edge of AI technology is and what possibilities it will bring to our lives and work, along with specific tool names.

Part 2: The reality of AI adoption – How is it changing the workplace?

Based on the latest technology trends, the session then moved to a more practical topic: "How is AI changing the workplace?" The speakers shared their real-life experiences on how AI is affecting the workplace and changing the way we work.

2-1. The reality of AI adoption in large companies: The first step is Google Gemini

Although there is a strong desire to introduce AI, there are many hurdles to overcome, especially in large companies, such as cost, security, and integration with existing systems.

Kobayashi suggested realistic implementation steps, saying, "Introducing a paid AI tool that costs 3,000 yen per month to all employees is a big decision for a company. For that reason, I recommend starting with a free, high-performance AI tool such as Google's Gemini."

Chaen agreed, adding, "Companies are broadly divided into those that use Google environments and those that use Microsoft environments, but a major benefit for Google users is that they can use many of Gemini's features at no additional cost. Security and cost are always important issues for companies, and Gemini lowers the barriers to adoption on both fronts."

Kojima said that many companies are taking a lean approach, in which they first conduct proof-of-concept experiments with a few departments or volunteer employees, and then roll out the technology to the entire company once the use cases and effects become clear.

2-2. Paradigm shift in work style: Freedom from "jobs you don't want to do"

All the speakers emphasized the importance of "top-down decision-making" as the key to successfully implementing AI.

"At a cutting-edge company like GMO, the management has a strong commitment to introducing AI. It is essential that the top management understands the potential of AI and actively promotes it," says Chaen. Although he predicts that it may take "around 20 years" for AI to become a commonplace part of use throughout Japan, he sees the change as progressing steadily. In particular, he expressed the view that AI will become a full standard once the current generation of people in their 50s and 60s retire and the digital native generation takes on core roles.

Kobayashi offered a more optimistic outlook: "Talent students are already adept at using AI, and an increasing number of students are thinking, 'I don't want to work for a company that bans AI.' In order to secure talented personnel, companies will likely be forced to accelerate the introduction of AI. It's possible that we'll see major changes in around 10 years, not 20 years."

Kojima described the changes in work style brought about by AI as "freedom from work you don't want to do." "For example, isn't it natural that AI will replace tasks that not everyone is willing to do, but that are necessary, like taking minutes of meetings? There was resistance when automated ticket gates were introduced in the past, but they are now commonplace. In the same way, AI will gradually become indispensable," said Kojima.

2-3. What departments and occupations are likely to adopt AI?

So, in what specific departments and occupations is AI adoption beginning to take hold?

Chaen said, "Large companies often focus their AI education first on the IT department, departments with an interest in new technologies, or potential leaders and executive employees. Developing AI-savvy personnel within the company will make subsequent developments smoother."

More specifically, Kojima cited "the product development department, customer support department, and marketing and sales department of the IT division." It appears that AI is being introduced in areas where it can directly contribute to business improvement, such as programming support, automating inquiries, streamlining creative production, and speeding up document creation.

2-4. The decisive difference between those who can use AI and those who can't

The session also featured specific examples of people who are using AI effectively.

Kobayashi gave the example of his own student life: "If we let AI transcribe and summarize the contents of lectures in real time, the time we would have spent taking notes can be used for other studies or activities. Whether or not we use AI makes a big difference in how we use our time."

Chaen shared another shocking story: "While I'm in a meeting, I have about five AI tools running at the same time in the background. They research information, create drafts of materials, take minutes... They are able to work several times faster than a human being," said Chaen.

These stories show that AI is not just a convenient tool, but can be a powerful weapon that dramatically improves individual productivity. Whether or not we can maximize its benefits depends on the awareness and skills of each individual.

Part 3: Mindset Required for an AI Society – To Survive in an Era of Change

As technology evolves and the way we work changes, what kind of mindset and skills are required of us humans? In the third part of the session, with the theme of "The mindset required in an AI society," guidelines for surviving this era of change were presented.

3-1. Keep your antennae up! Gathering information and having the courage to try things out

In the rapidly changing world of AI, it goes without saying that it is important to constantly keep up with new information.

Chaen cited three elements as the characteristics of successful people. "First, a mindset of 'I want to make things easier.' The desire to leave tedious tasks to AI and increase efficiency is what drives people to try out new tools. Next is 'stamina.' It's said that you can work five or ten times as much if you use AI, but to do that you need stamina. And finally, 'intellectual curiosity.' Constantly seeking out new information and mastering AI faster and deeper than anyone else is a major advantage in today's world," said Chaen.

Kobayashi spoke from a slightly different angle about the importance of "the ability to learn from others." "Humans have information and knowledge that AI does not yet have. In order to access that information, it may be surprisingly important to have the ability to be loved by others, to be educated, and even to care about your appearance," said Kobayashi. This is a unique perspective that shows that human attractiveness will increase relatively in the age of AI.

Kojima emphasized the importance of "information literacy," which is the ability to discern the truth from the abundance of information available. "AI can sometimes generate erroneous information, but even without that, the world is full of fake news and misleading information. It is essential not to blindly accept new technology or information, but to check primary sources or actually experience it for yourself," said Kojima.

3-2. Co-creation with AI: The true value of "humanity" and "creativity"

As AI becomes able to handle many tasks, the value of "things that only humans can do" will increase.

Chaen said, "Many processes such as product production and delivery can be made more efficient using AI. However, the ultimate act of selling, especially sales of high-value products where relationships are important, still requires human effort," pointing out the importance of human skills such as communication skills and trust-building.

"AI can memorize large amounts of information, but humans forget in order to adapt to new environments and generate new ideas. It is not always right to memorize everything," said Kojima, pointing out the value of the human thought process that AI does not have.

The speakers also agreed that to maximize AI's capabilities, humans need to have the "ability to ask questions," in other words, prompt engineering skills. Chaen said, "To write a good book, the author must first have a wealth of information and deep insight. If we want AI to produce high-quality output, humans need to provide it with high-quality input (instructions and information)."

3-3. What skills will be essential in the next five years?

Finally, the participants reaffirmed the skills and mindsets that will be particularly important over the next five years or so. To summarize, the following elements can be mentioned:

• Adaptability to change and eagerness to learn: Not being afraid of new technologies, and being proactive in learning and trying them out.
• Critical thinking and information literacy: The ability to examine the reliability of all information sources, including AI, and discern their essence.
• Communication and empathy: The ability to build and maintain human relationships is something that AI cannot replace.
• Creativity and problem-solving skills: The ability to use AI as a tool to create new value.
• Ethics and Responsibility: The ability to understand the ethical issues associated with the use of AI and act responsibly.

The age of AI presents us with new challenges, but at the same time it may also provide us with an opportunity to rediscover and hone uniquely human capabilities.

Part 4: Lively Q&A from the audience – A crossroads of knowledge woven by participants and speakers

At the end of the session, there was an active Q&A session with the audience. Real questions and concerns about AI were presented, and the speakers responded carefully from their respective expert knowledge.

4-1. [Q1] How should we address legal and ethical issues surrounding AI? (Questioner: Researcher in the humanities and social sciences)

"I feel that the advances in AI have not kept pace with legal and ethical issues in the humanities and social sciences. How should we address these issues as we move forward with product development?"

• Kojima: "Japan's legal framework for AI is just beginning. The first step for companies is to organize the data they hold and clarify what they want to protect and what they want to use with AI. After that, they need to formulate guidelines that take copyright and privacy into consideration and proceed with use while ensuring transparency. This may take time, but careful discussion and consensus building are essential."
• Kobayashi: "At an individual level, it is important to keep up with the latest information on laws and regulations and predict future trends. For example, news such as the partnership between SoftBank and OpenAI can provide some insight into the Japanese government's stance and future direction of regulations."
• Chaen-shi: "In the business world, there is a way of thinking that calculates risk in terms of expected value. You compare and weigh the potential legal risks, their probability of occurrence, and the benefits that can be gained, and then decide how much risk you are willing to take. Of course, it is also important to work with lawyers and other experts and obtain legal advice."

4-2. [Q2] How can students try out expensive, state-of-the-art AI tools? (Questioner: Student)

"Many of the latest AI tools are expensive and out of reach for students. Are there any good ways to use them?"

• Chaen-shi: "Internships at companies that are proactive in using AI will be a very good opportunity. My company (Digirise) also provides interns with an environment where they can freely use the latest AI tools. By utilizing the resources of a company, you can gain experience that would be difficult to gain on your own."
• Kobayashi: "Many AI tools offer free trial periods or discount coupons through influencers when they are first released. When a new tool is released, it's a good idea to actively seek out such information. Gamma also had such campaigns in its early days."

4-3. [Q3] How to deal with AI "hallucination" and improve information literacy? (Questioner: General participant)

"There is concern that AI will tell plausible lies (hallucinations), but this is not limited to AI. Don't people of older generations also need the ability to discern information?"

• Kojima: "As you say, hallucination is not a problem specific to AI, but a risk that lurks in all information. The important thing is to make it a habit to always check the reliability of the source of information. For example, when obtaining corporate information, I always try to look for primary information, such as checking the IR materials on the official website."
• Kobayashi: "When using AI to generate information, adding a prompt such as 'please cite the source' can help to reduce the risk of hallucination to some extent."
• Chaen-shi: "AI tools with fact-checking functions are also appearing. It is also effective to follow reliable information sources (such as experts and proven influencers) and compare multiple sources of information. Please refer to my information on X (laughs)."

Through these Q&A sessions, the multifaceted issues surrounding AI and practical approaches to addressing them became clear.

Conclusion: Enjoying changes in the AI era and designing the future

The passionate session, which lasted about an hour, concluded with powerful messages from the speakers.

• Chaen-shi: "My X (@chayen_AIPlease follow us on our Twitter and YouTube channel for the latest information on AI. If your company is interested in AI training or development, please feel free to contact us. We will also have a booth today!
• Kobayashi: "I will also be starting a YouTube channel in earnest in the future, where I plan to share information about how to live in the age of AI and the latest information. Please search for 'Aoibot.' Trickle is also actively recruiting interns, so if you want to be at the forefront of AI, please come and join us!"
• Kojima: "At Crafter, we provide a safe generative AI tool called 'Crew' for businesses. AI will become a familiar tool not only for professionals but also for general employees. Please try using AI, even if it's a free tool, and experience its potential."

If I had to summarise the key messages from this session into three, they would be:

1. The evolution of AI will never stop, so embrace change without fear.
2. AI is not a threat, but a tool that augments human capabilities.
3. What is most important in the age of AI is the desire to keep learning and the pursuit of uniquely human values.

We still don't know the full extent of how AI will affect our work and lives. However, one thing is certain: it is up to us, humans, to design a new future together with AI. We hope this session will inspire you to take the first step towards that goal. Let's enjoy the unknown world of AI and turn change into strength.