Sections
00:00 What we're covering today
00:27 1. FAA starts using AI to plan flights around Washington
01:24 2. OpenAI proposes shared AI standards without requiring launch reviews
02:14 3. Sixty organizations aim to expand AI access in underrepresented languages
03:08 4. Bessent says US-China AI talks will continue in Shenzhen
04:02 5. Climate Week debate asks whether AI's benefits justify its power use
04:50 Visit Hot Tea

Disclosure
Narration uses an AI-generated voice.

Transcript
Welcome to Hot Tea for Tuesday, September 22, 2026. Today, OpenAI proposes international standards, a language coalition sets a five-year goal, and U S-China talks get a planned follow-up. We'll also examine the debate over A I's energy use. First, a new tool is advising flight planners around Washington. This briefing uses an A I-generated voice.

FAA starts using A I to plan flights around Washington. The FAA began limited use of SMART around Washington, D.C., on September 21. The Department of Transportation says it draws on 200 data streams, including weather, flight plans and staffing. It forecasts congestion, then recommends changes to flight schedules. According to The Hill, traffic analysts use SMART to help make decisions about airspace. It doesn't replace the controllers who keep aircraft safely separated. For now, this is a limited rollout, and SMART only gives advice. FAA staff review every recommendation, and local leaders can reject it. The announcement doesn't show that SMART makes tickets cheaper or flights safer. Bad weather will be its next test. Delay minutes, rejected recommendations and comparisons with current planning methods could show whether it works. Before expanding the rollout, the FAA could show when SMART helped and when staff rejected its advice.

OpenAI proposes shared A I standards without requiring launch reviews. OpenAI called for U S-led international technical standards on September 21. Its proposal covers common evaluations, human oversight and incident reporting. The company says the standards would not require licenses or reviews before a model launches. Governments would decide whether to make them law. Reuters reports that Sam Altman is due to address the UN Security Council on Wednesday. OpenAI has proposed a policy, not an international rule. The company also says fully autonomous recursive self-improvement is not happening today. Its warning about losing control describes a possible future. Published drafts could reveal who sets the tests and who can challenge a lab's results. Governments would still need to act before any legal duties took effect.

Sixty organizations aim to expand A I access in underrepresented languages. The Gates Foundation announced a 60-organization coalition on September 21. Over five years, the coalition aims to help an estimated 3.4 billion people use A I in their own language and voice. Technology companies, researchers, governments and community organizations have signed on. They plan to work on open data, benchmarks and tools people can use. The foundation is an interested participant. This is a target population, not a count of people already served. The coalition says its structure and governance will be developed over the next year. Tests in each language and trials led by local communities could show whether people understand and can use the answers. Translation coverage alone is not enough. Communities also need a say in whether and how others use their data.

Bessent says U S-China A I talks will continue in Shenzhen. Treasury Secretary Scott Bessent said on September 21 that senior U S and Chinese officials would meet in Shenzhen in about two months. Reuters reported that he made the comments to CNBC. Bessent said both sides agreed to a formal A I dialogue and an incident communication line. The planned meeting follows the notification proposal covered yesterday. This is still Bessent's account of the talks. The report doesn't show that either side has tested an incident line or settled the notification rules. Bessent said they still need to discuss the protocols. A joint statement and a confirmed meeting date would show how firm the commitment is. Named contacts and a practice exercise would offer stronger evidence that an alert can reach someone able to respond.

Climate Week debate asks whether A I's benefits justify its power use. A I's energy demand drew attention at New York Climate Week on September 21. The Associated Press reported disagreement over whether A I's future benefits would outweigh the burden from data centers. Le Monde reported that a French Senate committee found a similar problem. Missing data makes those environmental benefits hard to measure, and A I's overall footprint remains uncertain. Forecasts of cleaner computing don't measure today's emissions. Even a useful A I application may consume more resources than it saves. Operators could report electricity, water use and emissions beside the measured savings from each application. Using the same boundaries and time periods would make the costs and savings easier to compare.

That is the signal before the noise. This briefing was produced from Hot Tea's verified daily edition. For the complete briefing and every source link, visit Hot Tea dot A I.