A project by Gerd Leonhard
Released: September 15, 2026. Last Updated:
I am presenting this initiative because we are no longer merely witnessing an AI revolution. We are navigating an AI Emergency - where autonomous power is scaling faster than human accountability. An emergency not just in terms of risks, safety and security, but in a multitude of ways, some more urgent than others but altogether representing a very serious challenge (see the list below).
I have been looking at AI-related issues for over a decade, delivering over 150 talks and keynotes on this topic, made many films on this subject, and published my own AI GPTs. It all got started with my 2016 book 'Technology vs Humanity' (read this review of how the key messages in that book held up in the last decade:) - back then, the subtite of the book was 'The coming clash of humans and machines' :)
I have to admit that - since I am far from being a doomer but more of a Protopian - I have been more optimistic about AI until now, and have not felt the urgent need to propose Emergency Actions until earlier this year. In the past few weeks, atarting with the Hugging Face incident, the existential threats of the race towards AGI / ASI as well as the accelerating negative externalities of AI are starting to pile up much faster than the promised benefits.
I therefore want to propose that the EU and other countries, nations, states and even counties and cities DECLARE AI EMERGENCY, just like we declared the pandemic emergency with COVID-19 back in 2021. We need special emergency rules and regulations to take effect within a very short time to at least try and staff off the immediate threats, and then we need time to address the subsequent challenges (see the list below).
Declaring AI EMERGENCY does not mean catastrophe is inevitable. It means the window for preventing catastrophe - and for shaping a good outcome - is closing much faster than our current response to prevent it. We are simply not doing nearly enough to ensure The Good Future and are thereby boosting existential risk.
First, we need dramatic, urgent and exceptional action on the existential threats that have emerged on the past 6 weeks, and then we need decisive, smart and wise policy responses to the many other challenges of the AI / Intelligence Revolution.
There is no other topic more urgent than this one. All else can wait.
IMMEDIATE AND EXISTENTIAL THREATS: My latest list of worries
1) TECHNICAL DISASTERS: Misalignment, over-reliance on AI agents, human carelessness or nefarious hacking attempts could to catastrophic failures such as global IT system collapses, financial market manipulations, control system failures, energy grid collapses etc. My estimated likelihood of AI-led disasters: 40% within 12 months. Dario Amodei — We Must Pace the Frontier How Would AI Actually Kill Us All? What to Know About the AI Doomsday Debate Or, it could be much more gradual but still end up with humanity 'timing out': Ron Bodkin on Twitter / X: How AI Might Kill Us All. Gradual but total disempowerment Must-Read (click to download the clean PDF version of Ron's tweet) 2) MONETIZATION PRESSURES AND GREED: The financial boom generated by AI-related investments leads to increasing carelessness and exponentially larger willingness to risk everything to stay on-top of the AI race. Below: Henry M. Paulson Jr. and Robert E. Rubin (via the Washington Post, paywalled, archive link here) : "We worry that perverse financial incentives are making it difficult to protect humanity from the dangers of AI" Opinion | We managed financial crises. AI is a risk of a different kind. “There’s no question that competition between companies causes them to take shortcuts on safety,” said Stuart Russell, a professor of AI at the University of California, Berkeley" Gift-link below Why the AI race has its creators fearing human extinction
2) Very low chances of a USA/China AGI treaty within the next 6 months: My estimated likelihood of such a treaty is 10%
"Global pacing will require cooperation with China, the autocratic country with by far the most advanced AI capabilities. We must not be naïve here: the geopolitical stakes are so high that there will likely be stark limits on what can be achieved, especially at first. If we greatly restrain our AI capabilities in the belief that China will do the same, and then China defects, AI could be so powerful that such a defection could lead to their geopolitical dominance. Therefore any agreement must either have ironclad verifiability, or must be limited enough that defection would not be militarily existential. I suspect that not only the US but also China will have these concerns and anxieties. We should approach any global pacing decision, especially in the near term, in such a way that protects the lead of the US and its allies…" We Must Pace the Frontier by Dario Amodei "When Presidents Donald Trump and Xi Jinping meet this month, we encourage the leaders to consider history and work toward an “ACT” — AI Cooperation Treaty. The U.S. and China are the only developers of frontier AI technology. Neither trusts the other. That absence of trust is, ironically, what may help the two nations protect themselves and the world" Henry M. Paulson Jr, Robert E. Rubin "China’s foreign ministry appeared to reject the warnings on Monday as fear-mongering, saying AI development was “a force for good”. “Spreading various threat narratives and engaging in confrontation and malicious competition will only disrupt the global governance process of artificial intelligence and does not serve the interests of any party,” said ministry spokesperson Guo Jiakun" FT Sept 14, 2026 (gift link) "“Well I say this,” Trump said Sunday during his visit to Ireland. “We’re leading China on AI. We’re the most sophisticated country in the world and frankly I want to keep it that way, because whoever wins AI wins. And we can put guardrails, and we can do this and that, but I think you have a lot of negative forces that are bringing it up, that shouldn’t be bringing it up, and they’re bringing up things that won’t happen" Politico Sept 14, 2026
3) Gradual but total disempowerment of Humans
WHY DECLARE EMERGENCY NOW?
- This is happening because there are literally Trillions of €££ at stake - the rush to monetize, own and dominate the 'Intelligence Revolution' has become the driving force of markets, investors and entrepreneurs.
- This is happening because the United States, as the driving force of innovation in technology (along with China, of course), has lost the plot in every possible way - and its AI policy is only of the most disastrous recent aberrations: Domination at all costs. A Manhattan Project for AI. A super-dangerous fusion of Silicon Valley techno-optimism and a truly mad Trump agenda.
climate emergency (hopefully with a much faster uptake). This action would clearly be driven by Europe, initially, and maybe South America, but ultimately it's a global undertaking (read more under solutions and people/orgs)
These will be described in more detail in the microsite, on my website and in my upcoming keynotes and videos.
My intention with this initiative is not to scare people or to boost the already abundant fear, but to rally everyone involved in this debate (industry leaders, VCs, startups, AI developers, scientists, activists, rights organisations, politicians and governments) to agree to Emergency Actions ASAP.
Neither doom nor boom, Protopia not Dystopia or Utopia
My narrative is not about doom or boom, for or against AI. It's about how to safely and collectively develop, deploy and use one of the most powerful technologies ever developed by humans: Artificial Intelligence. AI is here to stay; in fact it's quickly becoming the #1 driver of our economies. I often call this era the 'Intelligence Revolution' because it is truly on the scale of the Industrial Revolution (well, much beyond, in fact). There is no way for us to put this genie back into the box (truly) - so now it's about making sure it remains a servant rather than becomes the Master, and that we create sustainable frameworks to govern it
On the one hand, we have potential 'nirvana' / utopia scenarios promising abundance and dramatic gains for everyone. On the other hand, we have wildly dystopian and catastrophic scenarios essentially predicting humanity's demise if and when we develop AGI / ASI (general or super intelligence). The fear of AI is growing even faster then the power of the latest models that are rolled out constantly!
AI IS 'THE END OF BUSINESS AS USUAL', AND AGI IS THE END OF 'HUMAN AS WE KNOW IT' - and we must prepare, NOW.
WHY do I propose to declare an AI Emergency?
Simply put, we are not prepared, there is no good plan, we're already behind, and there is no global entity authorized to deal with this.
AI Capabilities
AI capabilities are moving at hyper-exponential speed.
Business
Business is moving at market speed.
Governments
Governments are moving at institutional speed.
Humanity
Humanity is moving at human, linear, organic speed.
These are the top10 issues that are currently insufficiently addressed, and thus compounding the AI Emergency:
- Safety, security, control and human alignment are not the top priority of frontier AI labs; while the world is accelerating towards AGI & RSI (Recursive Self Improvement) anything that is slowing things down is simply no longer an option
- Jobs, work and employment (education and training): We are totally unprepared for a world where most routines will be automated, and where work will no longer be the most important thing in our lives, and where work will no longer be the key economic factor
- Geopolitics: US-inspired techno-optimism paired with extreme capitalism and authoritarian, neolibertarian principles significantly amplifies AI / AGI dangers and results in an AI arms-race, darwinistic win/lose and zero-sum economic dystopias. For many venture capitalists that are betting on disruption, AI destroying democracy increasingly seems like an acceptable outcome.
- Human abdication, cognitive outsourcing (and then, surrender), reductionism, synthetic everything
- Erosion of democracy due to AI tools and human-like agents that manipulate and boost mistrust while fuelling disengagement, AGI-fuelled dictatorship scenarios where democracy is an obstacle to technocracy
- Boosting of techno-optimism, rationalism, EA, transhumanism
- There is currently no 'Mission Control' for AI: the UN has launched many laudable initiaves but lacks the power to enforce a binding treaty, due to security council veto rights and overall failure of global governance instruments
- We have entered the Fear Society where lack of hope drives younger generatives to desparate political conclusions (such as the rise of extreme right-wing parties), creating what I call 'the bad future syndrome'
- The AI arms-race - Military applications
My videos on AIEmergency
All latest videos here
01:49
YouTube
AI needs a red line, now: The #AIEmergency is here. Time for action.
AI could become humanity’s most powerful tool - or a terrible master. In this urgent keynote excerpt (see https://youtu.be/CoccfS28B7E?si=Ji78T0VBnuSvMIYP for the entire keynote) Also at https://www.gerd.tv/programs/2026-investment-perspectives-armenia-gerd-leonhard-keynote : INCLUDES SLIDES, Dubs etc - free access #futurist Gerd Leonhard explains why AI's safety, AGI regulation and an international non-proliferation agreement for superintelligence must become global priorities NOW - before m
12:15
YouTube
The AI Emergency explained: Gerd Leonhard Interview with News Armenia (edited for clarity)
This is an edited version of my Interview on News.am see https://news.am/en/video/1062603 on the current #artificialintelligence debates and AI risks see my resources at https://www.gerdfeed.com/search/AI+kill and my posts on x https://gerdleonhard.com/xaiemergency Watch on Gerd.TV: https://www.gerd.tv/programs/ai-emergency-interview-gerd-leonhard The AI Emergency — Balancing Tool and Threat In this discussion, I share my perspectives on why AI is an unavoidable leap, but one that carries
59:31
YouTube
AI and Humanity: the Next 5 Years: Why Human-Centric AI is the Key (Gerd Leonhard in Yerevan)
I went back to Yerevan, Armenia to talk about what the next five years could look like as AI accelerates faster than any technology we have ever seen. The #intelligencerevolution is here but so is the #aiemergency ! Thanks to @c@cubeInvestamor inviting me; more about the event here: https://conference.cubeinvest.am/en I share how I use AI for research, idea generation, and simple creative tasks. I explain why this “power tool” can boost productivity while also creating serious risks without
04:29
YouTube
THE AI EMERGENCY IS HERE — AND HUMANITY IS NOT READY | Gerd Leonhard on CNN-News18
Nope this is not a doomer thing. Get real. In this excerpt from a CNN-News18 debate held on Sept 14, 2026, I explain why AI’s greatest near-term danger is not (yet) an “evil syknet machine” but powerful, out-of-control, unaligned autonomous agents making catastrophic mistakes without context, accountability or human wisdom. AI must remain under human control. Powerful tools need powerful rules. Watch the whole thing here: https://www.youtube.com/watch?v=ZFsUYmzMy0M #ArtificialIntelligence #t
02:08
YouTube
What should AI do — and who decides? Powerful keynote excerpt by futurist Gerd Leonhard
NEW: Get exclusive content and early releases via my new www.gerd.tv platform (free access) In this short excerpt from my Sibos Frankfurt 2025 keynote see https://youtu.be/pmbIHyeA3-I?si=7rp4aDdM9_OC8LU0 I examine one of the most disturbing statements I’ve encountered about the possible trajectory of advanced AI, and explain why we need to take these warnings seriously - without surrendering to either AI hype or AI doom. AI is becoming an extraordinarily powerful new capability, but intellige
01:47
YouTube
The Rise of Digital Labour: When AI Starts Doing the Work (Keynote Excerpt)
AI is evolving from a tool we use into a new 'digital entity' that is capable of "digital labour" - software that can reason, decide and act with increasing autonomy. In this keynote excerpt from his Sept 11, 2026 keynote in Yerevan / Armenia, futurist and #publicthinker Gerd Leonhard explores the rapid shift from AI assistants to autonomous AI agents, and what it will mean for work, business and society. *** Watch the complete keynote video on GerdTV https://www.gerd.tv/programs/2026-invest
02:05
YouTube
The #AIEmergency: Gerd Leonhard's Take #1: Too much of a good thing can be a very bad thing
The AI Emergency is here. And this is not 'doomer talk'. In this first episode on what I am now calling #AIemergency I lay out why I think we're nearing emergency status as far as AGI 7 ASI is concerned. We have invented something extremely powerful and hyper-transformative but we don’t spend much time or resources on making sure that its side-effects are not detrimental. We have power-tools but almost no #telos ****watch more videos on AI here: https://www.gerd.tv/categories/ai AI is a onc
Current updates - monitored with AI:)
Sept 8: Bottom line: the evidence still points upward on major-incident risk, especially around autonomous cyber agents and systems taking consequential action outside intended boundaries. Existential risk remains much harder to quantify, but the gap between capability growth and credible control mechanisms is becoming a mainstream concern inside labs and governments.
Latest must-reads (more on the READ PAGE)
Top must watch (more on the video page)
14:28
YouTube
AI ethics and the future of safety | Dr. Roman Yampolskiy | TEDxMiami
As artificial intelligence grows more powerful, so do the risks. In this riveting talk, Dr. Roman Yampolskiy looks at the long-term implications of advanced artificial intelligence. His talk asks one of the most important questions of our time: Can we control what we create? Roman V. Yampolskiy is a tenured associate professor in the department of Computer Science and Engineering at the Speed School of Engineering, University of Louisville (2008-). He is the founding and current director of t
07:13
YouTube
Pause AI Development NOW
If the leaders of the major AI companies admit that they are losing control of their extremely dangerous technology, it is time for an immediate PAUSE on advanced AI development and a permanent BAN on superintelligence. -- Senator Bernie Sanders is the senior senator from Vermont. He is the longest-serving independent in U.S. congressional history and the ranking member of the Senate Committee on Health, Education, Labor, and Pensions (HELP).
04:42
YouTube
Mayor Mamdani Announces One-Year AI Moratorium for Students Through 8th Grade | AE1N
New York City Mayor Zohran Mamdani has unveiled the city’s first comprehensive AI policy for public schools, imposing a one-year moratorium on student-facing generative AI from 2-K through 8th grade. High schools will pilot five vetted AI programs under teacher supervision, while all high school students will receive AI literacy lessons. Officials say the policy prioritizes human connection, critical thinking and student safety. For more details, watch our story and subscribe to our channel, DRM
The Governance Chasm
In 2024, the European Union passed the AI Act — the world's first comprehensive legal framework for artificial intelligence. It was a historic achievement. It was also already out of date. Policy operates on timescales of years: consultations, amendments, elections, ratifications. Technical breakthroughs now arrive in weeks. The gap between what regulators understand and what engineers are building grows wider with every model release.
WELCOME TO