国立研究開発法人審議会　情報通信研究機構部会（第50回）

国立研究開発法人審議会　情報通信研究機構部会（第51回）

国立研究開発法人審議会（第23回）

第53回独立行政法人評価制度委員会　議事概要

California Governor Newsom has signed S.B. 524, a bill that begins the long process of regulating and imposing transparency on the growing problem of AI-written police reports. EFF supported this bill and has spent the last year vocally criticizing the companies pushing AI-generated police reports as a service. 

S.B.524 requires police to disclose, on the report, if it was used to fully or in part author a police report. Further, it bans vendors from selling or sharing the information a police agency provided to the AI. 

The bill is also significant because it required departments to retain all the various drafts of the report so that judges, defense attorneys, or auditors could readily see which portions of the final report were written by the officer and which portions were written by the computer. This creates major problems for police who use the most popular product in this space: Axon’s Draft One. By design, Draft One does not retain an edit log of who wrote what. Now, to stay in compliance with the law, police departments will either need Axon to change their product, or officers will have to take it upon themselves to go retain evidence of what each subsequent edit and draft of their report looked like. Or, police can drop Axon’s Draft One all together. 

EFF will continue to monitor whether departments are complying with this state law.

After Utah, California has become the second state to pass legislation that begins to address this problem. Because of the lack of transparency surrounding how police departments buy and deploy technology, it’s often hard to know if police departments are using AI to write reports, how the generative AI chooses to translate audio to a narrative, and which portions of reports are written by AI and which parts are written by the officers. EFF has written a guide to help you file public records requests that might shed light on your police department’s use of AI to write police reports. 

It’s still unclear if products like Draft One run afoul of record retention laws, and how AI-written police reports will impact the criminal justice system. We will need to consider more comprehensive regulation and perhaps even prohibition of this use of generative AI. But S.B. 524 is a good first step. We hope that more states will follow California and Utah’s lead and pass even stronger bills.

　終戦80年、しかも排外主義の政治勢力が伸長というこの夏、戦争を扱った番組から特に印象的なものを振り返る。　広島テレビ『テニアン　玉砕と原爆の島』は、広島・長崎への発進基地となったテニアン島で、米軍の攻撃下で日本人移民が集団自決した悲劇や、かつての滑走路が訓練場として再整備されている実態を伝えた。ＮＨＫ『ＢＳスペシャル　原爆裁判～被爆者と弁護士たちの闘い～』は、１９６３年の東京地裁の「原爆投下は国際法違反」という判決を、提起した岡本尚一弁護士や原告の被爆者５人の側から見つめ、..

　毎年９月１日になると、１９２３年の関東大震災で起きた大惨事を思う。わが身を守ることで必死だったはずの人々が、なぜ朝鮮人や中国人を殺そうなどという意思や力を持ったのか。「震災時のデマ」だけでは腑に落ちない。 　父・崔昌華 […]

　８月20日～22日、アフリカの開発を議論する国際会議である第９回アフリカ開発会議（ＴＩＣＡＤ９）が神奈川県横浜市で開催された。ＴＩＣＡＤは１９９３年から日本政府が主導し、３年ごとに国連や世界銀行、アフリカ連合委員会など […]

今夏以降、外国人の強制送還が急増　日本生まれ児童も送還の異常事態 　出入国在留管理庁（入管）が在日外国人を強制送還する事例が今夏以降急増している。今年５月発表の「国民の安全・安心のための不法滞在者ゼロプラン」を背景に、入 […]

「本件訴訟をいずれも棄却する」 　８月26日。東京高裁１０１号法廷での萩本修裁判長の判決読み上げに、ほぼ満席の傍聴席から「人権守れ！」との声が飛んだ。 　スリランカ人男性ナビーンさんと日本人のなおみさんは２０１６年に結婚 […]

In a letter sent to Microsoft at the end of last month, EFF and five other civil society organizations—Access Now, Amnesty International, Human Rights Watch, Fight for the Future, and 7amleh—called on the company to cease any further involvement in providing AI and cloud computing technologies for use in Israel’s ongoing genocide against Palestinians in the Gaza Strip.

EFF also sent updated letters to Google and Amazon renewing our calls for each company to respond to the serious concerns we raised with each of them last year about how they are fulfilling their respective human rights promises to the public. Neither Google nor Amazon has responded substantively. Amazon failed to even acknowledge our request, much less provide any transparency to the public. 

Microsoft Takes a Positive Step Against Surveillance

On September 25, Microsoft’s Vice Chair & President reported that the company had “ceased and disabled a set of services” provided to a unit within the Israel Ministry of Defense. The announcement followed an internal review at the company after The Guardian reported on August 6 that the IDF is using Azure for the storage of data files of phone calls obtained through broad or mass surveillance of civilians in Gaza and the West Bank.

This investigation by The Guardian, +972 Magazine, and Local Call also revealed the extent to which Israel’s military intelligence unit in question, Unit 8200, has used Microsoft’s Azure cloud infrastructure and AI technologies to process intercepted communications and power AI-driven targeting systems against Palestinians in Gaza and the West Bank—potentially facilitating war crimes and acts of genocide.

Microsoft’s actions are a positive step, and we urge its competitors Google and Amazon to, at the very least, do the same, rather than continuing to support and facilitate mass surveillance of Palestinians in Gaza and the West Bank.  

The Next Steps

But this must be the starting point, and not the end. Our joint letter therefore calls on Microsoft to provide clarity around:

1. What further steps Microsoft will take to suspend its business with the Israeli military and other government bodies where there is evidence indicating that business is contributing to grave human rights abuses and international crimes.
2. Whether Microsoft will commit to publishing the review findings in full, including the scope of the investigation, the specific entities and services under review, and measures Microsoft will take to address adverse human rights impacts related to its business with the Israeli military and other government bodies.
3. What steps Microsoft has taken to ensure that its current formal review thoroughly investigates the use of its technologies by the Israeli authorities, in light of the fact that the same law firm carried out the previous review and concluded that there was no evidence of use of Microsoft’s Azure and AI technologies to target or harm people in Gaza.
4. Whether Microsoft will conduct an additional human rights review, or incorporate a human rights lens to the current review.
5. Whether Microsoft has applied any limited access restrictions to its AI technologies used by the IDF and Israeli government to commit genocide and other international crimes. 
6. Whether Microsoft will evaluate the “high-impact and higher-risk uses” of its evolving AI technology deployed in conflict zones.
7. How Microsoft is planning to provide effective remedy, including reparations, to Palestinians affected by any contributions by the company to violations of human rights by Israel.

Microsoft’s announcement of an internal review and the suspension of some of its services is long overdue and much needed in addressing its potential complicity in human rights abuses. But it must not end here, and Microsoft should not be the only major technology company taking such action.  

EFF, Access Now, Amnesty International, Human Rights Watch, Fight for the Future, and 7amleh provided a deadline of October 10 for Microsoft to respond to the questions outlined in the letter. However, Microsoft is expected to send its written response by the end of the month, and we will publish the response once received.

Read the full letter to Microsoft here.

情報通信審議会　情報通信技術分科会　電波利用環境委員会　CISPR B作業班（第28回）配付資料