The OxyContin of the Future
Senator Collins saw the last catastrophe coming. The question her own warning poses now is whether Congress will govern the next one before it arrives.
Written With Claude Sonnet 5
In February 2002, at a Senate hearing few Americans noticed, a first-term senator from Maine said something that reads today like prophecy.
The hearing was titled “OxyContin: Balancing Risks and Benefits.” Senator Susan Collins had come with the problems of Washington County, Maine — a rural county Down East where a new prescription painkiller was tearing through communities that were already struggling. She questioned Purdue Pharma’s marketing of the drug. She called for a comprehensive response: prevention, physician training, treatment, law enforcement. And then she said this, in the official record:
“Let us not forget that other powerful drugs, some of them in development now, may become the OxyContin of the future if we do not learn from the lessons of the past couple of years and act on them today.”
Congress did not act on them that day. Fifteen years later, in October 2017, Collins sat in another hearing room — this time as a senior senator, the opioid epidemic now the deadliest drug crisis in American history — and gave voice to a frustration millions of Americans shared:
“We passed legislation, increased funding, we recognize that you have to focus on education and prevention, law enforcement, treatment and recovery, and yet we seem not to be making any progress that we need to make.”
Both statements are true, and together they tell the real story. Collins warned early. Congress eventually spent heavily — billions in grants, block grants, treatment programs, many of them secured by Collins herself as a senior appropriator. And still, more than 700,000 Americans have died of drug overdoses since the crisis began, hundreds of them each year in Maine.
How can both things be true — early warning, sustained spending, catastrophic failure? Answering that question honestly is the most important thing Americans can do right now. Because the next “OxyContin of the future” is not a painkiller. It is already in millions of workplaces, and Congress is once again deciding — mostly by not deciding — how it will be governed.
What Congress Never Examined
The conventional story of the opioid epidemic has villains, and they earned the role: a company that marketed an addictive drug deceptively, distributors who shipped pills by the millions, regulators who moved too slowly. All true, all documented in court.
But there is a layer beneath that story that Congress, across three decades, never seriously examined — even as its own expert bodies described it. In 2017, the National Academies of Sciences, Engineering, and Medicine put it plainly: while overprescribing was a vector of the epidemic, the root causes lay in “lack of economic opportunity, poor working conditions, and eroded social capital in depressed communities, accompanied by hopelessness and despair.”
Where did that despair come from? Some of it, the economic research now shows, came from Washington.
In October 2000, Congress granted Permanent Normal Trade Relations to China; China joined the World Trade Organization the following year. Economists David Autor, David Dorn, and Gordon Hanson later documented what followed — the “China Shock”: researchers estimate between roughly one and 2.4 million American manufacturing jobs lost, concentrated in communities that never recovered. And economists Justin Pierce and Peter Schott established something more disturbing: counties most exposed to that trade liberalization suffered statistically significant, persistent increases in fatal drug overdoses — concentrated among working-age people without college degrees. The trade policy and the death toll were causally linked.
Maine lived this. Its paper and forest-products industry — the economic backbone of rural towns for generations — shed thousands of jobs in the decade after 2001. In Washington County, the very county whose suffering Collins brought to that 2002 hearing, the Baileyville mill changed hands, shut down its paper operations in 2007, and idled its pulp operations in 2009, with layoffs cascading through loggers, truckers, and small businesses. The shocks never fully stopped: the Jay mill closed in 2023 with 230 layoffs, and in late 2025 Woodland Pulp announced a temporary shutdown affecting roughly a third of its workforce.
Into that terrain of lost work, lost identity, and lost community came a heavily marketed painkiller. The epidemic was not just a drug problem that policy failed to solve. It grew in soil that policy helped till — and for twenty years, the federal response treated the chemical while leaving the soil unexamined. Congress funded treatment for despair’s consequences while never reckoning with despair’s causes, some of which traced back to its own votes.
That is the lesson worth carrying forward. Not “Congress did nothing” — Congress did a great deal, much of it necessary. The lesson is that *reactive spending cannot compensate for unexamined causes and deferred rules.* Prevention money arrived after addiction was entrenched. Marketing restrictions arrived after the marketing had worked. And the economic dislocation that made whole regions vulnerable was never treated as part of the same story at all.
The Next Shock Is Cognitive
Artificial intelligence is now doing to cognitive work what trade liberalization did to manufacturing work — with two differences that should command Congress’s attention: it is faster, and it is everywhere.
The China Shock was geographically concentrated: mill towns, factory counties. AI is a general-purpose technology reaching into customer service, legal review, financial analysis, software development, translation, administration — work spread across every county in America, including the white-collar refuges where displaced manufacturing families sent their children. Early signals are measurable: hiring for entry-level software roles most exposed to AI has contracted sharply from its 2022 peak, and in 2025 more than 50,000 announced job cuts cited AI as a contributing factor.
None of this makes a catastrophe inevitable — and that is precisely the point. Economists Daron Acemoglu, David Autor, and Simon Johnson have shown that AI can be deployed in two directions: to *automate* human work, treating workers as costs to eliminate, or to *augment* it, extending what workers can do and raising the value of their judgment. Which direction prevails is not a law of nature. It is a policy choice — shaped by tax incentives, procurement, research funding, labor law, and the rules Congress writes or declines to write. Work, as the opioid epidemic taught at terrible cost, is not merely income. It is identity, structure, dignity, and social fabric — the very things whose destruction filled the graveyards of the last shock.
Congress at the Crossroads — Again
Where does Congress actually stand? The record is a mix of genuine effort and familiar drift.
To its credit: in July 2025, when a provision buried in the budget reconciliation bill would have banned states from regulating AI for ten years, the Senate stripped it by a vote of 99 to 1. Senator Collins was one of the amendment’s co-sponsors, alongside Senators Blackburn, Cantwell, and Markey — after 39 Maine state legislators publicly urged her to oppose the moratorium. It was a rare, overwhelming assertion that in the absence of federal rules, states must remain free to protect their citizens. Yet the idea did not die: within months, industry allies were pressing to revive federal preemption of state AI laws through other legislative vehicles. The ground Collins helped hold in July 2025 is contested ground still.
Beyond that vote, a bipartisan Senate working group has called for $32 billion a year in federal AI research investment. Bills to simply *measure* what AI is doing to jobs — the Workforce Transparency Act, the AWARE Act, the AI Workforce Impact Study Act — have been introduced but, as of this writing, not enacted. Collins herself has legislated where the harms are clearest to her: co-sponsoring the Protect Elections from Deceptive AI Act, co-authoring guidance to help election officials counter AI disinformation, backing penalties for stripping AI watermarks, and directing $45 million to a University of Maine complex integrating AI into rural health education.
Set that record against the 2002 template and the parallel is exact. Then as now: early hearings, sincere warnings, targeted bills at the visible edges — and no structural framework governing the thing itself before it reshapes millions of lives. The FDA approved; the DEA reacted; Congress appropriated; nobody governed upstream. Today: agencies improvise, states experiment under threat of preemption, Congress studies — and deployment races ahead.
Senator Collins is not the obstacle in this story. In many ways she is its best witness. She holds the receipts of what deferred action costs: she chairs the Appropriations Committee that funded the cleanup of the last catastrophe, sits on the health committee that heard its earliest warnings — her own among them — and serves on the Intelligence Committee now watching AI’s risks unfold. Few living legislators are better positioned to insist that this time, the questions come first. The question this article poses is not whether she cares. It is whether Congress — with her senior leadership — will treat her own 2002 warning as the standard: *learn from the lessons of the past and act on them today.*
An Election-Year Note, Stated Plainly
Maine’s Senate race is one of the most watched in the country. Senator Collins, seeking a sixth term, faces either Troy Jackson or Saundra Pelletierin in November. Moonshot Press does not endorse candidates, and this article is not a verdict on that race. It is an argument that AI governance belongs *in* that race — and in every Senate race — as a question citizens ask rather than a topic candidates choose.
So here are five questions for both candidates in Maine, and for every reader’s own senators, wherever you live:
1. Preemption: Will you oppose any federal moratorium that strips states of the power to regulate AI before federal protections exist?
2. Measurement: Will you support legislation requiring the government to track, publicly, how AI is affecting jobs — before displacement shows up as a crisis?
3. Direction: Will you support tax and research incentives that reward companies for using AI to augment workers rather than replace them?
4. The terrain: What is your plan for the communities — rural and urban — where economic shocks have historically turned into public-health catastrophes?
5. The lesson: What, specifically, did Congress get wrong in the twenty years after the first OxyContin warnings — and what will you do differently with AI?
Their answers — from the incumbent and the challenger alike — will tell you more than any advertisement.
What Citizens Can Do
This article is part of The People’s Commission on Technology and the American Future, Moonshot Press’s civic platform for exactly this purpose: helping citizens understand, monitor, and shape how technology is governed, rather than merely enduring the result. The Commission’s tools include position tracking (where your representatives actually stand — votes, sponsorships, statements — on AI and labor legislation), plain-language Citizen Briefs like the one accompanying this article, deliberation forums, and impact mapping showing how technological change is reaching your own community.
Start small and specific. Ask the five questions above at any town hall, forum, or constituent call — candidates answer what constituents repeatedly ask. Follow the companion Citizen Brief for the current status of the AI bills before Congress and how to comment on them. And if you live in Maine: both campaigns will hold public events between now and November 3. The 2002 hearing record shows what happens when good questions are asked and then set aside. This time, the follow-up is ours to do.
Twenty-four years ago, a senator from Maine warned about the OxyContin of the future. The future arrived. It is arriving again — and this time, the warning is on the record before the damage, with time to act on it.
About this story: All quotations from Senator Collins are drawn from official Senate hearing records (Senate HELP Committee, February 12, 2002 and October 5, 2017). Economic findings are attributed to the published research of Autor, Dorn & Hanson; Pierce & Schott; and Acemoglu, Autor & Johnson; the characterization of root causes quotes the National Academies of Sciences, Engineering, and Medicine (2017). Congressional actions verified against Senate records and reporting by CNN, NBC, Maine Public, and legal-policy analyses, current as of July 2026. Moonshot Press does not endorse candidates. AI assistance: background research, source verification support, and draft revision were AI-assisted under the Moonshot Press Constitution; editorial judgment, framing, and final responsibility are human. Corrections: corrections@moonshot.press.


