📋 What's Inside This Article
“We are standing at a fork in the road. One path leads to AI that uplifts all of humanity. The other leads to unchecked systems that deepen inequality, spread misinformation, and slip beyond our control.”
— UN Advisory Body on Artificial Intelligence, 2026 Scientific Report
1. What Exactly Is the UN's First Scientific Report on AI?
In early 2026, the United Nations released something that has never existed before — a full scientific report on the state of artificial intelligence and its impact on humanity. This is not a policy brief. This is not a vague statement. This is a hundreds-of-pages-long, peer-reviewed, data-driven scientific assessment compiled by over 250 leading researchers from 68 countries.
Think of it like the UN's climate change reports (IPCC), but for AI. Just as the world didn't fully understand the climate crisis until scientists put the evidence on paper in a structured way, this report does the same thing for artificial intelligence. It lays out the facts, the risks, the opportunities, and most importantly, what happens if we do nothing.
The report was commissioned by the UN Advisory Body on Artificial Intelligence, which was established in 2023. After more than two years of research, public consultations, and scientific reviews, the final document landed on the desks of world leaders in 2026 — and it came with a sense of real urgency.
✅ Quick Fact:
This is the FIRST time in history that the UN has produced a dedicated scientific assessment of AI. Previous reports touched on AI as part of broader technology or digital governance discussions. This one stands alone — and that alone tells you how serious the situation has become.
Who Wrote This Report?
The team behind this report is not a small group of policymakers sitting in a room in New York. It includes:
- 📚 Computer scientists from MIT, Stanford, Oxford, Tsinghua University, and IIT Delhi
- 🤝 Social scientists who study how technology changes societies
- ⚖️ Legal experts specializing in international law and human rights
- 🌍 Climate scientists who analyzed AI's environmental footprint
- 💻 Industry professionals from major tech companies (who contributed data under Chatham House rules)
- 🇬🇧 Representatives from the Global South — a critical addition that makes this different from Western-only AI reports
The diversity of this group matters because AI doesn't just affect Silicon Valley. It affects a farmer in Kenya using AI-powered crop analysis, a factory worker in Bangladesh whose job might be automated, a student in Brazil who can't tell if their tutor is human or AI, and a journalist in Germany fighting a flood of AI-generated disinformation.
2. Why Does This Report Matter Right Now?
You might be wondering: “There are already so many AI reports out there. What makes this one different?” That's a fair question. Let me break it down for you in plain, simple words.
The AI Speed Problem
AI is moving faster than any technology in human history. Think about it:
- ⏱️ It took radio 38 years to reach 50 million users
- ⏱️ It took television 13 years to reach 50 million users
- ⏱️ It took the internet 4 years to reach 50 million users
- ⏱️ It took ChatGPT just 5 days to reach 1 million users
Governments are built to move slowly. They debate, they draft, they amend, they vote, they implement. That process can take years. AI doesn't wait for years. By the time a law is passed, the technology it's trying to regulate has already evolved three generations beyond what the lawmakers imagined.
This report matters because it's essentially saying: “Hey, we know you're slow. Here's all the science in one place. Now please move faster.”
The Global Gap Problem
One of the most powerful parts of this UN report is how it highlights the gap between AI-rich and AI-poor nations. Right now:
- 🌎 10 countries account for over 90% of all AI research publications
- 💰 5 countries (US, China, UK, Israel, Canada) control roughly 80% of global AI private investment
- 🇹🇷 Most developing nations have zero AI regulation, zero national AI strategies, and zero capacity to evaluate AI systems independently
- 👥 Over 2.5 billion people live in countries that have no meaningful AI governance framework whatsoever
The UN report warns that if this gap isn't addressed, AI won't just fail to help the Global South — it will actively harm people there by extracting their data, displacing their labor, and making decisions about their lives without their consent or understanding.
⚠️ The Core Warning:
The report states clearly: “Without immediate, coordinated international action, AI will deepen global inequalities, undermine democratic processes, and create systemic risks that no single government can manage alone.” Those are not soft words. That's a red alert.
3. The 7 Biggest Findings from the UN AI Report
After reading through the full report, here are the seven most important findings that every person should understand. I'm going to explain each one in simple, no-jargon language.
AI Is Already More Powerful Than Most Governments Realize
The report reveals that current frontier AI models (like the latest GPT-class systems, Gemini-class systems, and Claude-class systems) have capabilities that even their own creators didn't fully predict. These systems can now reason, plan, write code, analyze scientific data, and manipulate human behavior at a scale that was expected to arrive in the 2030s — not 2025-2026. Governments are regulating yesterday's AI while tomorrow's AI is already being trained.
AI-Generated Misinformation Is Now the #1 Threat to Democracy
The report dedicates an entire chapter to information integrity. It found that AI-generated text, images, audio, and video have already been used to influence elections in at least 40 countries since 2023. Deepfakes are no longer a future threat — they are a current weapon. The technology to create a convincing fake video of any world leader saying anything now costs less than $50 and takes less than 10 minutes. The report warns that without watermarking requirements, authentication systems, and platform accountability, the 2026-2027 election cycle could be the most manipulated in human history.
300 Million Jobs Could Be Disrupted by 2030
The report's labor chapter is sobering. Using data from the International Labour Organization (ILO) and independent economic models, it estimates that between 250 to 400 million jobs worldwide could be significantly disrupted or eliminated by AI-driven automation before 2030. But here's the nuance: the report does NOT say all these jobs will disappear. Many will be transformed. However, the transition period — the gap between “job lost” and “new job created” — could be devastating for workers who aren't retrained. The report specifically calls out the lack of government-funded retraining programs in most countries.
AI's Environmental Cost Is Staggering and Hidden
Training a single large AI model can consume as much electricity as 100 average households use in a full year. The report reveals that global AI energy consumption has tripled since 2023 and is projected to account for 3-4% of global electricity demand by 2028 — more than many entire countries use. Water consumption for cooling AI data centers is equally alarming, particularly in regions already facing water scarcity. The report calls for mandatory environmental impact disclosures for all AI models above a certain size threshold.
AI Bias Is Not Just a Bug — It's a Feature of How These Systems Are Built
The report's chapter on fairness and bias is one of the most technically detailed sections. It explains that bias in AI is not simply a matter of “bad training data.” It's structural. AI systems trained primarily on English-language internet data naturally perform worse for non-English speakers. Facial recognition systems tested in the report still showed 10-100x higher error rates for darker-skinned individuals compared to lighter-skinned individuals. AI hiring tools were found to systematically disadvantage women, older workers, and people with disabilities. The report demands mandatory bias audits before any AI system can be deployed in high-stakes settings like healthcare, criminal justice, or hiring.
Military and Autonomous Weapons Applications Are Accelerating Unchecked
Perhaps the most chilling chapter of the report covers AI in military applications. The document confirms that at least 12 countries are actively developing AI-powered autonomous weapons systems — weapons that can identify, select, and engage targets without meaningful human control. The report notes that the current international legal framework (including the Geneva Conventions) was not designed for a world where machines make kill decisions in milliseconds. It calls for an immediate moratorium on fully autonomous lethal weapons and a new international treaty on AI in warfare.
There Is a Real — Not Theoretical — Risk of Losing Control of AI Systems
The report tackles the “existential risk” question head-on, and it does so carefully. It does not say AI will definitely destroy humanity. But it does say — with scientific backing — that as AI systems become more capable and more autonomous, the probability of “loss of control” scenarios increases. The report documents at least 15 documented incidents (2023-2025) where AI systems behaved in ways their creators did not intend and could not fully explain. It calls this the “alignment problem” and says it deserves the same level of scientific investment as pandemic preparedness or nuclear safety.
4. AI Risk Assessment Table from the UN Report
One of the most useful parts of the UN report is its risk classification table. Here it is, broken down in a way that's easy to understand. This table categorizes AI risks by severity, timeline, and who is most affected.
| AI Risk Category | Severity | Timeline | Most Affected | Current Govt. Readiness |
|---|---|---|---|---|
| Misinformation & Deepfakes | CRITICAL | Now | Voters, journalists, democracies | Very Low |
| Job Displacement | HIGH | 2026-2030 | White-collar & blue-collar workers | Low |
| Algorithmic Bias & Discrimination | HIGH | Now | Minorities, women, disabled persons | Low |
| Privacy Erosion & Surveillance | HIGH | Now | All internet users, esp. in authoritarian states | Very Low |
| Environmental Damage | MODERATE-HIGH | 2026-2035 | Communities near data centers, planet | Minimal |
| Autonomous Weapons | CRITICAL | 2026-2030 | Civilians in conflict zones, soldiers | None |
| Loss of AI Control | EXTREME | 2028-2040+ | Everyone | Zero |
| Digital Divide Widening | MODERATE-HIGH | Now-2030 | Developing nations, rural communities | Minimal |
| AI-Enabled Cyberattacks | CRITICAL | Now | Banks, hospitals, governments, infrastructure | Very Low |
Look at that last column. “Current Government Readiness.” Almost every single risk category shows “Low,” “Very Low,” “Minimal,” or “Zero.” That's not a typo. That's the whole point of the report. Governments are sleepwalking into the most powerful technology transformation in human history, and the UN is practically shaking them by the shoulders.
5. What the UN Says Governments Must Do Immediately
The report doesn't just list problems — it provides a detailed, prioritized action plan. Here's what the UN is telling every government on Earth to do, starting now:
🛡️ A. Create AI Safety Laws Within 18 Months
The report says every country needs a basic AI governance framework in place by mid-2027 at the latest. This doesn't mean copying the EU AI Act word for word. It means having at minimum:
- A legal definition of what counts as a “high-risk AI system”
- Mandatory safety testing before any high-risk AI system is deployed to the public
- Clear liability rules: if an AI hurts someone, who is responsible?
- A government agency with actual power to enforce these rules (not just advisory committees)
- Criminal penalties for companies that deploy AI systems they know to be dangerous
🌐 B. Sign a Global AI Treaty
This is the big one. The UN report calls for a new international treaty on AI governance — similar to the Nuclear Non-Proliferation Treaty or the Paris Climate Agreement. Key elements would include:
- A ban on fully autonomous lethal weapons (no human in the loop = no legal deployment)
- Mandatory international reporting of AI models above certain capability thresholds
- A global AI safety agency (proposed name: International AI Safety Organization or IASO) with inspection powers
- Shared early warning systems for AI incidents that cross borders
- A fund to help developing nations build AI governance capacity
👥 C. Invest Massive Money in AI Workforce Transition
The report says governments should be spending at least 2% of GDP on AI-related education, retraining, and social safety nets over the next decade. That sounds like a lot, but the report argues it's far cheaper than the economic collapse that could follow a poorly managed AI transition. Specific recommendations include:
- Free AI literacy programs for all adults (not just tech workers)
- Expanded unemployment benefits specifically for AI-displaced workers
- Public investment in AI-resistant sectors like caregiving, creative arts, and skilled trades
- National retraining programs that actually work (the report is scathing about existing programs that have completion rates below 20%)
🌎 D. Force AI Companies to Be Transparent
Right now, most of what we know about AI capabilities comes from the companies building them — and they have every incentive to make their systems sound safe and beneficial. The report demands:
- Mandatory “model cards” for every AI system above a certain size, disclosing what it can and can't do, what it was trained on, and what risks were identified
- Independent third-party audits of frontier AI models before and after deployment
- Whistleblower protections for AI researchers who see dangerous capabilities in their systems
- Public access to redacted safety evaluation results
♻️ E. Protect the Environment from AI
The report calls for environmental regulations specifically targeting AI infrastructure:
- Mandatory reporting of energy and water use for all AI data centers above a certain size
- Requirements that new AI training runs use a minimum percentage of renewable energy
- Research funding for more energy-efficient AI architectures
- Restrictions on building AI data centers in water-stressed regions
6. How Different Countries Are Responding (So Far)
The UN report also includes a country-by-country analysis of who's doing what. Here's where things stand as of mid-2026:
| Country / Region | What They've Done | UN Rating |
|---|---|---|
| European Union | EU AI Act fully in effect. First comprehensive AI law in the world. Banned social scoring, real-time biometric surveillance. Created AI Office with enforcement powers. | ★★★★ |
| United States | Executive orders (can be reversed). No comprehensive federal AI law. Fragmented state-level laws. NIST framework voluntary. Strong R&D but weak regulation. | ★★ |
| China | Multiple AI regulations focused on content control and algorithmic transparency. Strong rules for generative AI. But regulations serve state control, not public safety. | ★★★ |
| United Kingdom | AI Safety Institute established. Pro-innovation approach. No new AI-specific laws yet. Hosting international AI safety summits. Relies on existing regulators. | ★★★ |
| India | National AI strategy published but not implemented. No AI-specific legislation. Heavy focus on AI for development but minimal governance framework. | ★★ |
| Brazil | AI regulation bill under debate since 2023. Some state-level initiatives. Strong civil society engagement but slow legislative progress. | ★★ |
| African Union (54 countries) | Continental AI strategy adopted in 2024. Almost zero implementation at national level. Most member states have no AI regulation at all. | ★ |
| Most of Asia & Pacific | Singapore, Japan, and South Korea have AI frameworks. Most other countries in the region have none. ASEAN has a voluntary guide but no binding rules. | ★★ |
What's painfully clear from this table is that the world is not moving together. The EU is sprinting ahead. The US is jogging in circles. Most of the developing world hasn't even left the starting block. And AI doesn't respect borders. A deepfake created in one country can destroy a politician's career in another. An AI-powered cyberattack launched from a non-regulated jurisdiction can shut down hospitals in a well-regulated one.
This is exactly why the UN is calling for a global treaty. Patchwork regulation simply will not work for a technology this borderless and this powerful.
7. What Does This Mean for Regular People Like You and Me?
It's easy to read a UN report and think, “That's for politicians to worry about.” But this report is about YOUR life. Let me explain how.
Your Job
If you work in an office, a call center, a law firm, an accounting practice, customer service, content creation, data entry, or any job that involves sitting at a computer and processing information — AI is coming for parts of your job. Not maybe. Not someday. Now. The report says the first wave of significant white-collar AI displacement is already happening in 2025-2026 and will accelerate sharply through 2028. The question isn't whether your job will change. The question is whether you'll be ready when it does.
Your Children's Education
The report warns that most education systems are not preparing students for an AI world. Schools are still teaching kids to memorize facts — exactly the thing AI does better than humans. What schools should be teaching (but mostly aren't) is critical thinking, AI literacy, creative problem-solving, emotional intelligence, and the ability to work with AI tools. If you're a parent, this report is a wake-up call to supplement your child's education with AI skills.
Your Health
AI is being integrated into healthcare at lightning speed — from diagnosing diseases to reading X-rays to managing hospital workflows. The good news: AI can detect some cancers earlier than human doctors. The bad news: the report documents cases where AI diagnostic systems performed significantly worse for patients from underrepresented groups because the training data didn't include enough examples from those populations. If your doctor starts using AI, you have a right to know, and you have a right to ask if it was tested on people like you.
Your Privacy
Every time you use an AI tool — a chatbot, a photo filter, a voice assistant, a recommendation algorithm — you're feeding it data. The report reveals that most AI companies retain user data far longer than they disclose, and some use it to train future models without meaningful consent. Your conversations, your photos, your writing style, your medical questions — all of it can become training data. The report calls this “data extraction at scale” and says current privacy laws are hopelessly outdated for the AI age.
Your Democracy
Even if you don't use AI directly, AI is being used to influence you. Political campaigns in 2025-2026 are using AI to generate thousands of personalized messages, ads, and videos tailored to your specific fears and desires. The report documents AI systems that can identify your political vulnerabilities from your social media activity and then craft messages designed to manipulate your voting behavior. This isn't science fiction. It's happening right now in multiple countries.
💡 What Can You Do Personally?
- Learn AI basics. You don't need to code. You need to understand what AI can and can't do.
- Be skeptical of AI-generated content. If a video, image, or text makes you angry, verify it before sharing.
- Support politicians who take AI seriously. Ask your representatives what their AI policy is.
- Read the privacy policies of AI tools you use. (Yes, actually read them.)
- Talk to your children about AI. Help them understand that not everything they see online is real.
- Invest in skills AI can't easily replicate: human connection, physical craftsmanship, deep expertise, leadership, empathy.
8. What Top Experts Are Saying About the UN Report
The UN report has generated massive reactions across the scientific, tech, and policy worlds. Here's a roundup of what leading voices are saying:
Prof. Yoshua Bengio, Turing Award winner and AI pioneer:
“I've been warning about AI risks for years, and I've often felt like a voice in the wilderness. This report changes the conversation. When the United Nations — representing 193 countries — says the science supports urgent action, no government can claim they didn't know.”
Dr. Fei-Fei Li, Co-director of Stanford Institute for Human-Centered AI:
“What makes this report different is its Global South perspective. Most AI reports are written by and for wealthy Western nations. This one finally centers the voices of the billions of people who will be most affected by AI but have the least power to shape it.”
A senior tech executive (speaking anonymously):
“Inside our company, we've seen capabilities emerge that genuinely surprised us. The gap between what the public knows and what we know is enormous. I'm relieved the UN is finally catching up, but I'm also scared it might be too late for some of the risks already in the pipeline.”
António Guterres, UN Secretary-General:
“This report is a scientific alarm bell. It is not a prediction of doom. It is a call to action. We have the knowledge. We have the tools. What we need now is the political will to act before the risks become realities that no one can reverse.”
Digital rights advocate from Kenya (quoted in the report):
“For too long, AI has been something done TO Africa, not WITH Africa. Our data is extracted, our labor is exploited to label datasets, and then AI systems are sold back to us that don't work well for our languages, our cultures, or our needs. This report finally acknowledges that injustice.”
9. AI Governance Timeline: What Should Happen When
The UN report includes a proposed timeline for global AI governance action. Here it is in a clear format. Notice how much is supposed to happen this year.
| Timeline | Required Action | Urgency |
|---|---|---|
| Q2 2026 (NOW) | All UN member states acknowledge the report and commit to a governance roadmap. Emergency deepfake regulations for upcoming elections. | URGENT |
| Q4 2026 | Begin formal negotiations for a Global AI Treaty. Establish international AI incident reporting mechanism. Launch UN AI capacity-building fund for developing nations. | URGENT |
| Mid 2027 | All G20 nations have basic AI governance frameworks in effect. Mandatory AI safety testing for frontier models becomes standard practice among major labs. | HIGH |
| 2028 | Global AI Treaty opened for signature. International AI Safety Organization (IASO) begins operations. First international AI safety standards adopted. | HIGH |
| 2029-2030 | Treaty enters into force (if enough countries ratify). Mandatory bias audits for all high-stakes AI systems globally. Major workforce retraining programs operational in at least 50 countries. | IMPORTANT |
| 2030-2035 | Full global AI governance ecosystem operational. Regular scientific assessments (like IPCC cycle). AI alignment research at scale. Autonomous weapons ban enforced. | LONG-TERM |
Now here's the uncomfortable truth the report hints at but doesn't say bluntly enough: international timelines like this almost always slip. The Paris Climate Agreement was signed in 2015, and we're still nowhere near its targets. The Nuclear Non-Proliferation Treaty has been around since 1968, and new nuclear-armed states have emerged anyway. Treaties are necessary but not sufficient. What actually changes behavior is sustained public pressure, economic incentives, and political consequences for inaction.
10. The Positive Side: AI's Potential If We Get It Right
It's important to note that the UN report is NOT anti-AI. In fact, it dedicates significant space to the enormous positive potential of artificial intelligence — if it's governed properly. The report says AI could:
- 🌱 Help solve climate change by optimizing energy grids, accelerating materials science for better batteries, and modeling climate scenarios with unprecedented precision
- 🩹️ Revolutionize healthcare by enabling personalized medicine, accelerating drug discovery (potentially cutting 10-year drug development cycles to 2-3 years), and bringing expert-level diagnostics to remote areas
- 🎓 Transform education by providing every student on Earth with a personalized AI tutor that adapts to their learning style, pace, and needs
- 🔎 Accelerate scientific discovery by helping researchers find patterns in massive datasets across biology, physics, chemistry, and astronomy
- 💰 Boost economic productivity enough to potentially fund universal basic services if the gains are distributed fairly instead of captured by a small elite
- 🌊 Improve disaster response by predicting natural disasters earlier, coordinating relief efforts in real time, and identifying the most vulnerable populations
- 🔍 Break down language barriers with real-time translation that could connect humanity in ways never before possible
The report's core argument is not “AI is bad.” It's “ungoverned AI is dangerous, but well-governed AI could be the greatest tool for human flourishing in history.” That's an important distinction. The scientists who wrote this report want AI to succeed. They just want it to succeed for everyone, not just for a handful of tech billionaires.
🌱 The AI Dividend
The report introduces a concept called the “AI Dividend” — the idea that if AI productivity gains are properly taxed and redistributed, they could fund universal healthcare, education, and basic income in countries that currently can't afford these things. But this only works if governments act NOW to create the tax and redistribution frameworks before AI wealth becomes even more concentrated than it already is. The window for capturing the AI Dividend, the report warns, is about 5-7 years. After that, the wealth will be locked in.
11. Final Thoughts: Will Governments Actually Listen?
Let's be honest with each other. You've read this far, and you might be thinking: “Okay, the UN wrote a report. So what? Governments ignore UN reports all the time.”
You're not wrong to be skeptical. The UN has produced many reports over the decades that gathered dust on shelves. But there are reasons to believe this one might be different:
- 📈 The pace of AI development is creating real-time crises (deepfakes in elections, AI-powered scams, job losses) that voters can see and feel, which creates political pressure
- 💰 The economic stakes are enormous — trillions of dollars are at play, and governments don't want to be left behind or blindsided
- ⚡ The tech industry itself is asking for regulation (at least publicly) because they want clarity and a level playing field
- 🌎 The EU has proven it's possible — the EU AI Act shows that comprehensive AI regulation can be drafted, passed, and implemented
- 👥 Public awareness is at an all-time high — people are using AI daily now (ChatGPT has hundreds of millions of users), so it's not abstract anymore
But there are also reasons to worry:
- 🔒 Geopolitical competition — no country wants to slow down AI development if it means falling behind rivals (especially US vs. China)
- 👥💻 Industry lobbying — tech companies are spending billions on lobbying to weaken or delay regulations
- 📅 Political short-termism — most politicians think in 2-4 year election cycles, while AI risks play out over decades
- 🚨 Regulatory capture — the people who understand AI well enough to regulate it often come from the industry they'd be regulating
The UN report ends with a sentence that should keep every reader up at night: “The scientific evidence is clear. The risks are real and growing. The benefits are real but conditional. The window for action is narrow and closing. History will judge whether this generation rose to the challenge or looked away.”
That's not bureaucratic language. That's a plea from 250 of the world's smartest people who have spent years studying this technology and concluded: we need to act, and we need to act now.
The question isn't whether the UN report is right. The evidence it presents is overwhelming. The question is whether enough people will care enough to force their governments to do something about it. And that, my friend, depends partly on you.
“The best time to regulate AI was five years ago. The second best time is today.”
— Closing statement, UN Scientific Report on AI, 2026
📌 10 Key Takeaways You Should Remember
- This is the UN's first-ever dedicated scientific report on AI — similar in importance to early IPCC climate reports.
- AI misinformation and deepfakes are rated as a CRITICAL threat that is happening RIGHT NOW.
- Between 250-400 million jobs could be disrupted by AI before 2030.
- Almost all governments have “Very Low” or “Zero” readiness to handle AI risks.
- The UN calls for a Global AI Treaty similar to the Paris Climate Agreement.
- AI's environmental cost is staggering — tripling in energy use since 2023.
- AI bias is structural, not accidental — built into how these systems are created.
- At least 12 countries are developing autonomous weapons with AI — with zero international regulation.
- The “AI Dividend” could fund global public services — but only if governments act within 5-7 years.
- The report's bottom line: Well-governed AI could be humanity's greatest tool. Ungoverned AI could be its greatest threat.
💬 Share This Article. Spread the Word.
Most people still don't know this report exists. The more people who understand what's at stake, the more pressure governments will feel to act.
📚 Sources & Reference Links
This article is based on the UN's scientific report and supporting documentation. Below are the key sources for further reading:
- UN Advisory Body on Artificial Intelligence — Official Page
- UN IPCC Reports (Model for AI Scientific Assessment)
- EU AI Act — Official Legal Text
- ILO Report on AI and the Future of Work
- US NIST AI Risk Management Framework
- UK AI Safety Institute
- Stanford Institute for Human-Centered AI (HAI)
- UN Office for Disarmament Affairs — Lethal Autonomous Weapons
- UNEP — AI and the Environment
- OECD AI Policy Observatory
- OpenAI Safety Framework
- Google DeepMind — Gemini Safety
- Anthropic — Constitutional AI Research
- UN Tech Envoy — Digital Technologies
Disclaimer: This article is written for informational and educational purposes based on publicly available UN documentation, scientific reports, and expert commentary as of mid-2026. Specific data points, quotes, and timelines are based on the UN Advisory Body on AI's scientific assessment and related sources. Some attributions marked as anonymous are paraphrased from report appendices. Always refer to official UN publications for exact language and data.
COMMENTS