Tokyo AI Strategy Council
https://www.youtube.com/live/opvFauAVeEU
1st 2024-12-05
takahiroanno It was the Tokyo AI Strategy Conference. we were doing a live stream on YouTube. There were many helpful comments. Do it openly → Lots of people will see it → Wisdom will be gathered
It's nice, isn't it?
https://gyazo.com/7679491d275b1575ea8d8f05d299a081
miyasaka Thank you for today. I would also like to thank everyone who participated, especially those who provided helpful wisdom in their comments. It was a rare opportunity for the administration to run an online conference in this style, and I was very excited. After the meeting was over, the staff members were very happy with the results. https://gyazo.com/b0171a15751dc4134e944ea48fc77b00
1:06:00
https://www.youtube.com/live/9OFWcoZ2XhM
o3-mini-high.iconSummary (concise to the point).
The Tokyo Metropolitan Government held the second meeting for the "AI Strategy Development".
The three perspectives of "the power to use," "the power to listen," and "the power to create" plus governance development are presented as a summary of the issues related to the administrative use of AI.
Discussions included the use of AI in a wide range of fields, including education and healthcare, the design of incentives for TMG employees, and how to attract foreign talent.
Deputy Governor Miyasaka's assertion.
I want to be a municipality that is proactive in using new technology in the government sector. We will need someone to ask about AI strategies."
He stressed the need to create a mechanism (staff incentives and consultation system) that allows the field to actively adopt AI as well as the governance and rules associated with its use.
Opinions that came up frequently in the discussion.
Catching up on the latest AI trends is limited to one institution or a few people.
Sharing examples and know-how in an open forum allows for flexibility in dealing with risk.
2. when introducing AI, it is necessary to specify not only "efficiency improvement" but also improvement of service quality and benefits to citizens and employees.
Since resistance may arise in the field if AI is perceived as taking away work or leading to personnel reductions, we will present measures to "redirect the amount of time saved by AI utilization to higher value-added work".
In the field of education, "utilization of large-scale language models such as ChatGPT" has great potential, but if not properly adopted by instructors, there are concerns about "decline in thinking ability" and "misuse," so discussions proceeded with reference to cases in Taiwan and the Khan Academy.
4. how to develop governance (transparency, fairness, and safety) in the extensive use of AI and data.
The issue of the demarcation point of responsibility, such as when an AI answers incorrectly.
Management of study data, unification of disclosure criteria, human intervention at risk (safety design), etc.
Revised draft of 4 principles for TMG AI development.
Adjusted wording to include "based on the assumption that all agencies will be deployed" to "proceed quickly with a view to common use in the future.
It includes risk assessment and review functions and agile review at the time of use.
The state of the data center.
Awareness of server expansion and power issues due to demand for AI development.
He pointed out the need for backup in the event of a disaster and the need for wide-area collaboration based on geographic location.
Future Directions.
A draft of the AI strategy will be compiled and further brushed up by summer.
We will seek a prompt and transparent governance structure while focusing on priority areas such as improving staff literacy, working with overseas human resources and big tech, and the education sector.
---
This page is auto-translated from /nishio/東京都AI戦略会議 using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.