AIが期間内の動向を整理
OpenAI clarifies safety disclosure rules, makes ad creation conversational, and Google advances voice and weather AI
The AI news on September 17 centered on OpenAI clarifying safety disclosure rules, moving ad creation into chat, improving pricing and image features, and Google bringing voice APIs and weather forecasting closer to practical use. Anthropic also published its monthly misuse report, making this a day when both usability and safety advanced for business users.
SOURCE CHECK
Primary Sources 72
Primary Sources
SOURCE CHECK 1
x.com
SOURCE CHECK 2
openai.com
SOURCE CHECK 3
openai.com
SOURCE CHECK 4
openai.com
SOURCE CHECK 5
x.com
SOURCE CHECK 6
ai.google.dev
SOURCE CHECK 7
ai.google.dev
SOURCE CHECK 8
ai.google.dev
SOURCE CHECK 9
blog.google
SOURCE CHECK 10
blog.google
SOURCE CHECK 11
blog.google
SOURCE CHECK 12
blog.google
SOURCE CHECK 13
blog.google
SOURCE CHECK 14
blog.google
SOURCE CHECK 15
ai.google.dev
SOURCE CHECK 16
blog.google
SOURCE CHECK 17
blog.google
SOURCE CHECK 18
blog.google
SOURCE CHECK 19
blog.google
SOURCE CHECK 20
blog.google
SOURCE CHECK 21
blog.google
SOURCE CHECK 22
ai.google.dev
SOURCE CHECK 23
blog.google
SOURCE CHECK 24
blog.google
SOURCE CHECK 25
blog.google
SOURCE CHECK 26
blog.google
SOURCE CHECK 27
blog.google
SOURCE CHECK 28
developers.googleblog.com
SOURCE CHECK 29
blog.google
SOURCE CHECK 30
blog.google
SOURCE CHECK 31
blog.google
SOURCE CHECK 32
blog.google
SOURCE CHECK 33
blog.google
SOURCE CHECK 34
blog.google
SOURCE CHECK 35
blog.google
SOURCE CHECK 36
blog.google
SOURCE CHECK 37
blog.google
SOURCE CHECK 38
developers.googleblog.com
SOURCE CHECK 39
developers.googleblog.com
SOURCE CHECK 40
blog.google
SOURCE CHECK 41
blog.google
SOURCE CHECK 42
blog.google
SOURCE CHECK 43
blog.google
SOURCE CHECK 44
blog.google
SOURCE CHECK 45
blog.google
SOURCE CHECK 46
x.com
SOURCE CHECK 47
ai.google.dev
SOURCE CHECK 48
ai.google.dev
SOURCE CHECK 49
blog.google
SOURCE CHECK 50
deepmind.google
SOURCE CHECK 51
ai.google.dev
SOURCE CHECK 52
ai.google.dev
SOURCE CHECK 53
blog.google
SOURCE CHECK 54
x.com
SOURCE CHECK 55
blog.google
SOURCE CHECK 56
blog.google
SOURCE CHECK 57
blog.google
SOURCE CHECK 58
blog.google
SOURCE CHECK 59
ai.google.dev
SOURCE CHECK 60
ai.google.dev
SOURCE CHECK 61
ai.google.dev
SOURCE CHECK 62
ai.google.dev
SOURCE CHECK 63
ai.google.dev
SOURCE CHECK 64
blog.google
SOURCE CHECK 65
openai.com
SOURCE CHECK 66
deploymentsafety.openai.com
SOURCE CHECK 67
help-lb.openai.com
SOURCE CHECK 68
x.com
SOURCE CHECK 69
anthropic.com
SOURCE CHECK 70
anthropic.com
SOURCE CHECK 71
blog.google
SOURCE CHECK 72
x.com
Key Points
- 1OpenAI published criteria for tracking and disclosing misalignment cases
- 2Chat-based ad creation and management became easier in ChatGPT Ads
- 3GPT-5.6 pricing updates and Fast mode make cost optimization easier
- 4Google strengthened the practical value of Live API for voice and WeatherNext 3
- 5Anthropic published its September misuse report, adding more safety reference material
OpenAI clarifies disclosure criteria for safety issues
OpenAI published a framework for tracking, investigating, and disclosing model misalignment, along with six cases from the past six months. By organizing what should be made public and when, companies can more easily carry out safety checks and explain risk management internally.
Ad creation and image generation move closer to business workflows
ChatGPT Ads now makes it easier to create ads through conversation, and a test offering for Sponsored Agents has begun. With Shopify and HubSpot integrations added, the flow from creation to delivery and measurement is easier to connect. ChatGPT Images 2.5 also helps reduce the number of revision cycles.
Pricing updates make AI operations cheaper
GPT-5.6 Luna and Terra were lowered in price, and Sol gained a Fast mode. Since the changes also apply to ChatGPT Work and Codex, teams can run the same work at lower cost. This is especially relevant for teams with high usage volumes.
Google advances voice AI and weather forecasting
Gemini 3.8 Live and Extended Thinking reached GA, making it easier to choose between low-latency conversation and deeper reasoning in Live API. WeatherNext 3 was also introduced with improved precipitation forecasting. These are useful signals for companies building voice apps or managing weather-dependent work.
Anthropic’s monthly report makes safety review easier
Anthropic published its September misuse report, organizing abuse cases and the responses to them. It confirmed multiple malicious attempts and updated its detection and blocking efforts. That makes official safety information easier to review when evaluating adoption.