While Washington and Beijing have long sparred over the hypothetical dangers of artificial intelligence yet to come, a quieter reckoning is being demanded: the harms are not approaching — they have already landed. Algorithmic bias, deepfakes, automated injustice, and AI-fueled disinformation are not warnings on the horizon but conditions already shaping millions of lives. The argument being advanced is that two rival superpowers, so consumed by the race to govern tomorrow's AI, have neglected the urgent work of repairing today's damage — and that this neglect, shared equally by both nations, m
U.S. and China Must Address AI Damage Now, Not Future Threats
The window for prevention has closed.
Mark
So the piece is saying we should stop worrying about AI going wrong and start fixing what's already broken?
Mimi
Not exactly. It's saying the focus on preventing future catastrophes has crowded out attention to damage that's happening right now. Deepfakes, algorithmic bias in hiring, misinformation—these are real harms affecting real people today.
Luke
But the source material is extremely thin. We get the argument, but no examples of specific harms, no numbers, no evidence that this damage is actually widespread or irreversible.
Mimi
That's fair. The piece is making a philosophical case more than a factual one. It's saying the premise of prevention has already failed, but it doesn't prove that premise with concrete data.
Mark
Why would China and the US cooperate on this? They're competitors in AI development.
Mimi
The argument is that they both have citizens harmed by current systems, so they share an interest in remediation. It's not about abandoning competition—it's about finding common ground on fixing present damage.
Luke
But we don't know if either government actually sees it that way, or if there's any real momentum toward that kind of cooperation. The piece is prescriptive, not descriptive.
Mark
What would cooperation actually look like?
Mimi
The piece suggests things like joint research on detecting AI misinformation, sharing auditing practices, maybe compensation for people harmed, transparency standards.
Luke
Again, those are possibilities the writer is floating. There's no reporting here on whether anyone is actually pursuing these things.
Mark
So it's an opinion piece making an argument, not reporting on something that's happening.
Mimi
Exactly. It's a call for a shift in how both countries approach AI policy—from prevention to remediation.
Luke
And the confidence level is medium, which makes sense. The core claim—that current AI damage is irreversible and requires repair—is asserted but not substantiated.
Il Polso
- The window for prevention has closed — AI harms are not theoretical but already embedded in hiring systems, content feeds, surveillance infrastructure, and electoral discourse.
- Both the United States and China have poured diplomatic energy into regulating AI systems that don't yet exist, while existing systems cause unaddressed, compounding injury to real people.
- The tension is not merely technical but moral: deepfake victims, algorithmically rejected job seekers, and disinformation targets are accumulating without recourse or remedy.
- A reframing is being proposed — shifting the policy conversation from strategic competition over future AI supremacy to shared remediation of present-day damage.
- Concrete cooperation could take the form of joint misinformation detection, bias auditing standards, transparency requirements, and compensation mechanisms for those already harmed.
- The piece lands as a challenge to governments and tech companies alike: stop rehearsing futures that may never arrive and begin accounting for the present that already has.
While Washington and Beijing have long sparred over the hypothetical dangers of artificial intelligence yet to come, a quieter reckoning is being demanded: the harms are not approaching — they have already landed. Algorithmic bias, deepfakes, automated injustice, and AI-fueled disinformation are not warnings on the horizon but conditions already shaping millions of lives. The argument being advanced is that two rival superpowers, so consumed by the race to govern tomorrow's AI, have neglected the urgent work of repairing today's damage — and that this neglect, shared equally by both nations, may itself be the most consequential policy failure of the age.
The debate over artificial intelligence between Washington and Beijing has long been oriented toward the future — toward hypothetical dangers, unbuilt systems, and speculative catastrophes. But a pointed argument is now being made that this forward-looking anxiety has come at a steep cost: the harms that AI is producing today, in systems already deployed, are accumulating without adequate remedy.
The indictment is specific. Deepfakes are circulating. Algorithmic bias is embedded in hiring, lending, and criminal justice. AI-powered disinformation is shaping elections. Automated screening tools are rejecting job applicants through processes no one fully understands. These are not warnings — they are conditions. And the window for prevention, the argument goes, has already closed.
What remains is remediation. And here lies the unexpected proposal: that the United States and China, despite their strategic rivalry, share a concrete interest in addressing the injuries their own populations are already experiencing. A worker displaced by automation faces the same fundamental problem regardless of which flag flies over the algorithm that replaced them. A person whose likeness appears in a deepfake suffers the same harm regardless of where the model was trained.
Cooperation need not mean harmony on all fronts. It might mean joint research into detecting AI-generated misinformation, shared standards for auditing algorithmic systems, or transparency requirements that let people know when AI is making decisions about them. Some damage, the piece acknowledges, cannot be fully undone — only managed and learned from.
The deeper provocation is about where attention goes. Governments and companies have invested enormously in imagining AI risks that may never materialize. The call being made is to redirect that same intensity toward harms that are already measurable, already present, and already shaping millions of lives — to stop waiting for a future crisis and begin accounting for the one already here.
The argument over artificial intelligence between Washington and Beijing has long fixated on what comes next—the hypothetical dangers lurking in systems not yet built, the theoretical harms that advanced AI might inflict if left unchecked. But this forward-looking anxiety, the thinking goes, has allowed both nations to overlook something more pressing: the damage that artificial intelligence is doing right now, today, in the systems already deployed and already reshaping how people live.
The case being made is straightforward and unsettling. The United States and China have spent considerable diplomatic and policy energy debating how to prevent future AI catastrophes—establishing guardrails for systems that don't yet exist, negotiating frameworks for technologies still on the drawing board. Meanwhile, the harms that current AI systems are producing—in hiring decisions, in content moderation, in surveillance, in the spread of misinformation—are accumulating without adequate remedy. These are not speculative injuries. They are happening now, to real people, in real institutions, and they are largely unrepaired.
The premise underlying this critique is that prevention, as a strategy, has already failed. The damage is not coming; it has arrived. Deepfakes are circulating. Algorithmic bias is embedded in consequential systems. Disinformation campaigns powered by AI are influencing elections and public discourse. Job applicants are being rejected by automated screening tools that no one fully understands. The window for prevention has closed. What remains is the work of remediation—figuring out how to undo harm, how to compensate those who have been wronged, how to rebuild trust in institutions that have deployed these tools without adequate safeguards.
This reframing carries significant implications for how the two superpowers might approach their relationship around technology. Rather than treating AI governance as a problem to be solved through competing regulatory frameworks and strategic competition over who builds the most advanced systems first, the argument suggests that both nations have a shared interest in addressing the concrete injuries their populations are already experiencing. A Chinese worker displaced by automation faces the same fundamental problem as an American worker. A person whose image has been used without consent in a deepfake video suffers harm regardless of which country's AI system created it.
The proposal, then, is not naive. It does not suggest that the United States and China will suddenly abandon their strategic competition or that they will agree on everything. Rather, it suggests that on the specific question of AI's present-day harms, cooperation might be not just possible but necessary. Both nations have citizens harmed by current AI systems. Both have institutions struggling to manage the consequences. Both have incentives to develop solutions that work, that are seen as legitimate, and that restore some measure of confidence in how these technologies are being used.
What this would look like in practice remains unclear. It might involve joint research into how to detect and counter AI-generated misinformation. It might mean sharing best practices for auditing algorithmic systems for bias. It might require establishing compensation mechanisms for people who have been harmed by AI systems, or creating transparency standards that allow people to understand when and how AI is making decisions about them. It might mean acknowledging that some of the damage already done cannot be fully repaired, only managed and learned from.
The deeper challenge is one of attention and priority. Governments and tech companies have invested enormous resources in thinking about AI risks that may or may not materialize. The question being posed is whether that same intensity of focus might be redirected toward harms that are already manifest, already measurable, already affecting millions of people. It is a call to stop waiting for the future to arrive and to start dealing with the present that is already here.
Vuoi la storia completa?
Leggi l'originale su The New York Times ↗
Technology