Obama tells Jeffries: make AI the Democrats' first fight
The pacing debate left the labs on Saturday. By Sunday it had a party.
Barack Obama is telling Democrats to claim AI policy as a 2028 issue, and The New York Times has the transcript. Speaking at a private fundraising event, the former president said of the technology: "This is something that is moving very fast in private hands, and if we don't get on top of it, I think can be dangerous." He urged House Minority Leader Hakeem Jeffries, per a partial transcript released by Obama's office, to build "a framework for a very public conversation" on AI if Democrats retake the House in the November 3 midterms, and called on 2028 candidates to put safety and economic disruption at the center of their platforms. The intervention matters because it lands days after Anthropic's researcher warnings pushed the topic from conference halls into congressional hearings — and because it splits the Democratic Party's usual posture: Obama is asking for a framework, not a pause, while Senate negotiators are drafting a duty-of-care bill that could let Washington block model releases. The practical read: AI regulation just became a floor-fighting issue for a party that spent 2023-2025 afraid of looking luddite in tech donor circles. We tracked the industry side of this shift yesterday — Altman matches Amodei's evaluator pledge, Musk says 'Dario is right'.
LG has answered the spy-TV allegations, and the answer is a parse of the word "recording." The company calls reports that its televisions capture and upload audio "misleading," saying Automatic Content Recognition, voice recognition, and interest-based advertising are optional and off by default, and that ACR fingerprints audio through the TV's internal audio processor rather than the microphone — so, in its telling, no recordings are collected. What the statement skips: independent analysis showed the sets scanning home networks for phones and other devices even when the TV appears offline, and LG did not address how much data it collects in total, who receives it, or how its consent menus are designed — the same pattern we covered when the fine print made you liable for anyone its TV's AI mic records. Denials that answer a narrower question than the one asked are the standard playbook here; the network-scanning finding went unaddressed for a reason.
What to watch: whether Jeffries' camp adopts the framework language before the midterms, and whether any Democratic candidate picks it up as a 2028 plank.
Should AI regulation be a party platform issue, or a bipartisan technical one? Tell us in the comments.
