⏱️ Lectura: 9 min
More than 200 tech entrepreneurs and scientists signed an open letter to the UN this week asking, in their words, to control artificial intelligence technology “for the good of humanity.” The letter arrives amid a wave of crossfire statements about AI risks: from Nvidia CEO Jensen Huang questioning the figures used by the most alarmist sectors, to US congressmen claiming those risks are no longer science fiction.
📑 En este artículo
The episode sums up a year in which much of the industry went from resisting any regulation to publicly asking someone to set limits. For anyone following the AI race closely, understanding who says what and why is no longer optional.
TL;DR
- Hundreds of tech leaders signed a letter to the UN asking to control AI development for the good of humanity.
- Nvidia CEO Jensen Huang called a key figure circulating in the AI risk debate unscientific.
- Congressman Ted Lieu warned that AI risks no longer belong to the realm of science fiction.
- NPR published a guide to understanding the different factions debating AI safety in 2026.
- A White House science advisor publicly discussed the guardrails being evaluated for AI systems.
- Universities like Indiana University are strengthening cybersecurity against the risks AI introduces into research.
- A report warned that AI’s productivity gains in science are reduced by the time spent validating results.
What happened: the AI risk debate reaches the UN
According to Associated Press, the letter gathers signatures from founders, researchers, and executives asking the United Nations to control the development of artificial intelligence before the race between labs becomes irreversible. The text doesn’t detail a specific legal mechanism, but it raises the tone: it no longer talks about “principles” but about direct control.
Almost simultaneously, Jensen Huang came out to question one of the most repeated figures in the debate. According to Yahoo Finance, the Nvidia CEO stated that this number, used to gauge the scale of AI risk, has no scientific basis. Huang didn’t offer an alternative figure, but his statement exposes the tension between those who build the models and those who warn about their consequences.
The counterpoint came from Congress. Representative Ted Lieu told Politico that AI risks “no longer belong to the realm of science fiction,” and asked his colleagues to treat the issue with the same urgency as other technological risks that have been regulated for decades.
Context and history: from private warnings to the public letter
The debate over the risks of artificial intelligence isn’t new, but its tone changed in 2026. For years, the strongest warnings came from independent researchers or letters signed behind closed doors; now the UN and the US Congress discuss them publicly, with names and titles.
Weeks earlier, the UN had already taken a step in that direction by creating a scientific panel dedicated to artificial intelligence, inspired by the climate IPCC model. The new letter from tech leaders adds pressure from the private sector, the same sector that until recently explicitly asked regulators not to touch open models.
To make sense of the noise, NPR published a guide that groups the debate’s participants into factions with different interests and diagnoses. It’s a useful map because it explains why the same question (how much risk is there?) gets such opposing answers depending on who answers it.
The factions in the AI risk debate
Simplifying NPR’s guide, the debate can be read as four blocks. None of them is monolithic (there are nuances within each), but the table helps place each public statement in context.
| Faction | Core diagnosis | Recent example | What they propose |
|---|---|---|---|
| Existential risk | Advanced AI could escape human control | Signatories of the UN letter | International treaties and compute limits |
| Accelerationists | The real risk is falling behind, not AI itself | Sector funding unrestricted compute | Less regulation, more investment |
| Figure skeptics | Risk numbers lack solid scientific basis | Jensen Huang | Evidence-based debate, not projections |
| Institutional regulation | The risk is real and requires legal frameworks now | Ted Lieu and other lawmakers | Laws and oversight from Congress |
How to follow the debate with primary sources
Following this debate in real time doesn’t require waiting for a media outlet to summarize it. The letter’s text, bills, and safety papers can be checked directly with simple tools.
To track which AI-related bills are moving through the US Congress, the public congress.gov API lets you search by keyword:
curl "https://api.congress.gov/v3/bill?query=artificial+intelligence&api_key=$CONGRESS_API_KEY" \
-H "Accept: application/json"The response includes the bill number, its status, and the lawmakers cosponsoring it: the kind of data that backs up statements like Ted Lieu’s.
For the academic side of the debate, the one that feeds both those who talk about existential risk and the skeptics, arXiv exposes a REST API with no authentication required:
import feedparser
feed = feedparser.parse(
"http://export.arxiv.org/api/query?"
"search_query=cat:cs.CY+AND+abs:AI+safety"
"&sortBy=submittedDate&sortOrder=descending&max_results=10"
)
for entry in feed.entries:
print(entry.title, entry.published)The entry.published field gives the exact date: if the paper someone cites on social media is more than a year old, it’s probably already been superseded by more recent work.
💡 Tip: before sharing an AI risk figure, look up the original paper or report with these two tools. Most viral figures lose their methodological context along the way.
Impact and analysis
The letter to the UN contrasts with the stance much of the same industry held months earlier, when it asked Washington not to restrict open-weight models. The contradiction isn’t necessarily hypocrisy: regulating the development of frontier systems and curbing the distribution of open models are different asks, even though both use the words “safety” and “risk”.
Universities like Indiana University are already reacting in practice: their cybersecurity teams are strengthening research defenses against the risks AI itself introduces into the scientific process, from contaminated synthetic data to agents that run code without supervision.
A THE Journal report found that the productivity gains AI brings to science are cut back by the time researchers must spend validating each result. That finding is, at its core, the same tension as the political debate: AI’s promise and its risk grow together, and validation costs real time.
📌 Note: none of the actors cited has published an independently verifiable risk probability figure. The statements from Huang and the letter’s signatories are, so far, qualitative positions.
What’s next
In the short term, the letter to the UN doesn’t obligate anything: it’s a declaration of intent, not a treaty. Its real weight will be measured by whether the General Assembly gives it a place on its agenda and whether the existing AI scientific panel takes it as input. In the US Congress, the question is whether statements like Ted Lieu’s translate into concrete bills or remain hearings without a vote.
flowchart TD
A["Tech leaders' letter"] --> B["UN General Assembly"]
B --> C["AI scientific panel"]
D["Ted Lieu and Congress"] --> E["AI bills"]
C --> F["Possible international framework"]
E --> FOn the technical side, a White House science advisor has already publicly revealed what type of guardrails the administration is evaluating, though without committing to a timeline. Meanwhile, every new frontier model that gets released will reopen the same question: who decides how much risk is acceptable.
Try it yourself: read the full text of the letter to the UN or run the arXiv API query above and draw your own conclusions about how solid the evidence is behind each side.
📖 Summary on Telegram: View summary
Frequently Asked Questions
What exactly does the tech leaders’ letter ask the UN?
It asks the United Nations to control the development of artificial intelligence “for the good of humanity,” according to the text cited by Associated Press, though it doesn’t propose a specific legal mechanism.
What number did Jensen Huang question?
Huang didn’t specify the exact figure, but according to Yahoo Finance, he called a number circulating in the public debate over AI risks unscientific.
What did Ted Lieu say about AI risks?
The congressman told Politico that these risks no longer belong to the realm of science fiction and called for treating them with legislative urgency.
Does the UN already have an AI panel?
Yes, the United Nations created a scientific panel on artificial intelligence weeks earlier, inspired by the climate IPCC model, which could serve as a channel for the new letter.
How can I follow the debate without relying on summaries?
With the public congress.gov and arXiv APIs shown in this article, you can track bills and safety papers directly from the source.
Does this affect open source models?
Not directly: the letter talks about AI development in general, while the debate over open models, which much of the industry asked not to restrict, is a separate discussion.
References
- Associated Press via MSN: text of the tech leaders’ letter to the UN.
- Yahoo Finance: Jensen Huang’s statements on the figures in the AI risk debate.
- Politico: interview with Congressman Ted Lieu about AI risks.
- NPR: guide to the factions in the AI safety debate.
- News at IU: how Indiana University is strengthening cybersecurity against AI risks in research.
- THE Journal: report on how validating results reduces AI’s productivity gains in science.
📱 Enjoy this content? Follow @programacion on Telegram for daily tech content in Spanish: quick summaries, fresh content every day.
0 Comments