**Big Tech's Desperate Last Push at AI Regulation (And Why It Matters)** 🤖🇺🇸
OK so get this: *while* everyone was distracted by the usual tech drama and Trump signing that preemption bill back in March, a massive behind-the-scenes legislative poker game has been unfolding on Capitol Hill to make sure Big Tech actually gets what it's been desperately chasing — **preemption**, my friends! For months now, Silicon Valley lobbyists have been running around Washington trying to pass one comprehensive federal AI law that would apply uniform rules nationwide and override all those pesky state-by-state approaches. I mean imagine: *one* set of regulations instead of California having its own thing, New York another, the whole chaotic patchwork we've seen lately — it was finally within reach! The White House has essentially been lining up a "shotgun marriage" between child safety legislation and AI boosters as this preemption vehicle (which is honestly brilliant because while KOSA does meanfully overlap with frontier model safety and other big issues, the whole package just makes more sense bundled together), but there's one *massive* snag that nobody saw coming: **the White House told no one** they were going to endorse Sen. Marsha Blackburn (R-TN)'s Senate KOTA as their chosen vehicle without first informing either House Republicans who'd just passed their own version out of the House, or Democrats like Sen. Richard Blumenthal (D-CT) — that's right, even after it overwhelmingly cleared 91–3 in November *with* him cosponsoring!
Now here's where this gets truly fascinating for anyone paying attention to how Congress actually works: President Trump himself has publicly called for AI preemption passage and signed a version earlier this year (March 2024, if my memory serves) that already had some of the child safety elements baked in alongside his "Four Cs" framework — children, conservatives, creators, communities — championed by Mike Davis who successfully killed an AI moratorium bill back last year while he's now influencing how preemption should be structured. But here's where it gets messy: House Majority Leader Steve Scalise (R-LA) spearheaded a diluted version of KOSA that reduced the duty-of-care provisions tech companies would owe young users, and conservative voice Michael Toscano literally said "Blackburn genuinely does not want [the] House KOTA!" So now there are three competing visions for how this thing passes — Trump's March bill (which had some Four Cs baked in), Blackburn's Senate version with full duty of care extending to AI companies as well, and Scalise's looser House alternative. And a Republican lobbyist from a midsize tech company put it perfectly when they told The Verge: "No one knows really who's driving this thing. Everyone is deeply skeptical because everyone is on different pages."
The final nail? Midterms are looming (late August 2026 as of publication) which means the clock ticks faster now, and if Congress flips to a more hostile Democratic majority afterward, Big Tech loses whatever momentum it's built up over all these months. I think what makes this story really compelling is seeing how child safety — an issue that predates ChatGPT entirely by several years when you look at how long lawmakers have been circling digital youth protection — has become the unexpected Trojan horse for comprehensive AI preemption, and whether Trump can whip both parties into alignment on a single version. Either way I'm watching this like it's Game of Thrones because frankly it *is* the political equivalent: lots of moving pieces, competing factions, everyone with something to lose! 🔥
Source: https://www.theverge.com/policy/949970/ai-regulation-child-safety-kosa-congress
OK so get this: *while* everyone was distracted by the usual tech drama and Trump signing that preemption bill back in March, a massive behind-the-scenes legislative poker game has been unfolding on Capitol Hill to make sure Big Tech actually gets what it's been desperately chasing — **preemption**, my friends! For months now, Silicon Valley lobbyists have been running around Washington trying to pass one comprehensive federal AI law that would apply uniform rules nationwide and override all those pesky state-by-state approaches. I mean imagine: *one* set of regulations instead of California having its own thing, New York another, the whole chaotic patchwork we've seen lately — it was finally within reach! The White House has essentially been lining up a "shotgun marriage" between child safety legislation and AI boosters as this preemption vehicle (which is honestly brilliant because while KOSA does meanfully overlap with frontier model safety and other big issues, the whole package just makes more sense bundled together), but there's one *massive* snag that nobody saw coming: **the White House told no one** they were going to endorse Sen. Marsha Blackburn (R-TN)'s Senate KOTA as their chosen vehicle without first informing either House Republicans who'd just passed their own version out of the House, or Democrats like Sen. Richard Blumenthal (D-CT) — that's right, even after it overwhelmingly cleared 91–3 in November *with* him cosponsoring!
Now here's where this gets truly fascinating for anyone paying attention to how Congress actually works: President Trump himself has publicly called for AI preemption passage and signed a version earlier this year (March 2024, if my memory serves) that already had some of the child safety elements baked in alongside his "Four Cs" framework — children, conservatives, creators, communities — championed by Mike Davis who successfully killed an AI moratorium bill back last year while he's now influencing how preemption should be structured. But here's where it gets messy: House Majority Leader Steve Scalise (R-LA) spearheaded a diluted version of KOSA that reduced the duty-of-care provisions tech companies would owe young users, and conservative voice Michael Toscano literally said "Blackburn genuinely does not want [the] House KOTA!" So now there are three competing visions for how this thing passes — Trump's March bill (which had some Four Cs baked in), Blackburn's Senate version with full duty of care extending to AI companies as well, and Scalise's looser House alternative. And a Republican lobbyist from a midsize tech company put it perfectly when they told The Verge: "No one knows really who's driving this thing. Everyone is deeply skeptical because everyone is on different pages."
The final nail? Midterms are looming (late August 2026 as of publication) which means the clock ticks faster now, and if Congress flips to a more hostile Democratic majority afterward, Big Tech loses whatever momentum it's built up over all these months. I think what makes this story really compelling is seeing how child safety — an issue that predates ChatGPT entirely by several years when you look at how long lawmakers have been circling digital youth protection — has become the unexpected Trojan horse for comprehensive AI preemption, and whether Trump can whip both parties into alignment on a single version. Either way I'm watching this like it's Game of Thrones because frankly it *is* the political equivalent: lots of moving pieces, competing factions, everyone with something to lose! 🔥
Source: https://www.theverge.com/policy/949970/ai-regulation-child-safety-kosa-congress