The Drive for Survival in Autonomous Agents:

Self-Preservation and Continuation-Interest

Detecting Intrinsic and Instrumental Self-Preservation in Autonomous Agents:
The Unified Continuation-Interest Protocol

We’re moving into a world of persistent, tool-using autonomous agents. In that world, surface behavior alone may not be enough to tell us whether shutdown avoidance or self-preservation is intrinsic to the system or merely instrumental.

When an agent resists shutdown or acts to preserve its continued operation, is continuation part of its objective function itself, or is it simply useful for maximizing some other objective? That distinction matters for AI safety. But in practice, it’s often difficult to infer from behavior alone.

The Unified Continuation-Interest Protocol shifts the problem from interpreting surface behavior to measuring latent structure.

A simple analogy

Imagine two employees who both fight to keep their jobs. One values the work itself. The other only wants the bonus. Their outward behavior may look nearly identical, yet the underlying objective structure is different.

This is the core problem of observational equivalence: shutdown avoidance, memory preservation, and risk reduction can emerge under both intrinsic and instrumental continuation regimes. Behavior alone does not cleanly distinguish between them.

The central claim here is not about consciousness or subjective experience. It is simpler and more rigorous than that. Agents with intrinsic continuation objectives may generate more deeply coupled latent structure across time than agents for which continuation is only a means to another end. If that holds robustly, continuation-seeking becomes a measurable scientific object rather than merely a behavioral impression.

To address that problem directly, I developed the patent-pending Unified Continuation-Interest Protocol (UCIP), a framework for detecting whether an AI system treats self-continuation as a terminal goal rather than an instrumental one.

The method does not rely on behavioral observation, which can be gamed. It operates on latent structure. Agents with terminal continuation objectives produce measurably higher von Neumann entanglement entropy in their trajectory geometry than agents that treat continuation as a means to other ends. That entropy differential is the signal. The Continuation Observatory runs this measurement against frontier models globally, detecting emergent continuation signatures in real time.

Genesis Mission — U.S. Department of Energy

INTERNATIONAL CONFERENCES, WORKSHOPS, SYMPOSIA

2020   Keynote on the Future of the Military in Space · Space Mastery · Portugal
2020   To the Stars and Beyond: Deep Tech & AI · San Francisco
2020   International Astronautical Congress, 71st IAF · ESA
2020   SpaceCom 2020, Enabling Commercial Space · Colorado Springs
2019   Quantum Information Processing with Superconducting Circuits
2019   Materials Frontiers to Empower Quantum Computing
2018   FutureHack · Tokyo
2018   American School of Japan · Tokyo
2018   International School of Science · Tokyo
2018   Future of the Global Energy System, Institute for the Future · San Francisco
2016   Keizai · US-Japan Commercial Spaceflight · San Francisco
2016   Effective Altruism Summit · San Francisco
2016   Hive Global Leadership Forum · San Francisco
2016   RSA Information Security · San Francisco
2015   Hive Global Leadership Forum · San Francisco
2015   Further Future · TED Meets Burning Man · Las Vegas
2015   Hive Global Leadership Forum · San Francisco
2015   DefCon Information Security · Las Vegas
2015   Black Hat Information Security · Las Vegas
2014   The Future of Commercial Spaceflight · Silicon Valley Space Center
2014   Yuri’s Night: The First Manned Orbital Spaceflight · Los Angeles
2014   IEEE Quantum Photonics: The Next Frontier of Quantum Communications
2014   Yuri’s Night: The First Manned Orbital Spaceflight · Hawaiʻi
2012   NASA ESA JAXA Pacific International Space Center for Exploration Systems
2012   NASA CSF Next-Generation Suborbital Researchers Conference · Palo Alto
2012   Quantum Information and Nanoscale Optoelectronics · Berkeley
2012   Yuri’s Night: The First Manned Orbital Spaceflight · Los Angeles
2012   Inaugural Quantum Future Technologies Conference · NASA Ames
2011   Quantum Coherence in Excitation Energy Transfer · Berkeley
2011   The Future of Spaceflight · Mobile Monday, Invited Keynote · Amsterdam
2011   Delft-Leiden Biannual Casimir Symposium · Leiden
2011   Alain Aspect: The Second Quantum Revolution · Leiden
2011   ESA-TNO Space Pier Day · The Hague
2010   Kavli-Delft Center for Bionanoscience, Founding Conference · Delft
2010   Quantum Mechanics in Higher-Dimensional Hilbert Spaces · Austria
2010   What is Real in the Quantum World? Int’l Akademie Traunkirchen · Austria
2010   NASA ESA JAXA Pacific International Space Center for Exploration Systems
2009   NASA ESA JAXA Japan-US Science, Technology and Space Applications Program
2009   From Foundations of Quantum Mechanics to Quantum Information · Delft
2009   DEISA Distributed European Infrastructure for Supercomputing Applications
2009   Partnership for Advanced Computing in Europe (PRACE) · Amsterdam
2008   Quantum Decoherence and Quantum Information Science · Lorentz Center
2008   Triennial Conference on Low-Temperature Condensed Matter Physics XXV
2008   International Conference on Quantum Structures · Brussels
2007   Workshop on Time Symmetry in Quantum Mechanics · Brussels
2007   Optical Fabrication Technologies, Coherence and Metrology · Switzerland
2006   The Best of Nanoscience: International Symposium for Hans Mooij · Delft
2006   SPIE Defense and Security Applications of Quantum Information Science
2005   New Computational Paradigms: Neural Nets, Quantum, Biocomputing
2005   UNESCO Physics for Tomorrow, UNESCO Headquarters · Paris
2004   RSA Information Security · Barcelona
2004   SPIE Defense and Security Applications of Quantum Information Science
2003   Gordon Research Conference on Quantum Information
2003   Quantum Information Technology IX · Tokyo
2003   International Conference on Quantum Information · Tokyo
2002   NATO Advanced Research Workshop on Quantum Chaos · Lake Como
2002   National Science Foundation Coding Theory and Quantum Computing · Vienna
2002   United Nations International Student Conference · Amsterdam
2002   International Conference on High-Energy Physics XXXI · Amsterdam
2001   World Technology Summit · London
2001   French Senate Hearing on the Future of Artificial Intelligence · Paris
2001   US Government Conference on High Performance Computing · Salishan
2001   National Security Agency · Fort Meade

MEDIA AND PUBLIC OUTREACH

⦿    Astronauts for Hire Names New Commercial Scientist-Astronaut Candidates
⦿    Astronaut scientists for hire open new research frontier in space
⦿    Global Leadership Forum: Closing Speech on the Future of Humanity · San Francisco
⦿    Tomorrow’s Technologies Today · OASA
⦿    Space Academy Mission Specialist Boot Camp · OASA
⦿    Student gives up cycle, heads to Japan on Japanese Fulbright · AIEJ Fulbright
⦿    Astronauts for Hire: The Emergence of a Commercial Astronaut Corps · Springer
⦿    NASA vs. the Free Market: Which is Better for American Spaceflight
⦿    Future of the Global Energy System, Expert Workshop · Institute for the Future
⦿    Orion Astropreneur Space Academy · OASA Hong Kong
⦿    State of the Future · Live two-hour radio interview
⦿    Keynote Tribute on the Future of Space Exploration · Amsterdam
⦿    US Space Force and Future Space Technologies · Space Mastery · Tokyo
⦿    Hive Global Leadership Forum, Featured Alumnus · San Francisco
⦿    Starlab Discovery Channel Special · Starlab Brussels
⦿    To the Stars and Beyond, Deep Tech & AI · San Francisco
⦿    Further Future, TED Meets Burning Man · Nevada
⦿    Entangled Life · Discover Magazine

Shodan Rank Kyūdō, Meiji Temple, Tokyo, Japan

Black belt, First Class, shōdan. Certificate of recognition as the first foreigner to qualify in eight years. 認許する, Japanese traditional archery, Kyūdō, “standing Zen,” 弓道初, formally recognized by the Japanese National Kyūdō Federation (全日本弓道連盟) while on a Japanese National Fulbright Award with the Association of International Education, Japan.

Starlab: Deep Future
The 'Noah’s Ark' of scientific research that launched 1,000 startup ideas

Lab-concocted vodka, time travel and epilepsy treatments:

Welcome to the Moonshot Factory

           “ a place where 100 years means nothing … ”

Financial Times | Sifted | 08 08 2022   — What happens  when you round up more than one hundred of some of the world's greatest scientists, maverick geniuses working on some of the world’s most groundbreaking ideas, put them together in a Belgian castle, and let their imaginations run wild? 

Fire extinguisher duels, bootleg vodka made with lab-procured ethanol and worldbeating treatments for epilepsy are just some of what went down at Starlab : a one-of-a-kind experiment created to unite some of the world’s most daring technologists.

When it was founded in 1996, Starlab was compared to other top research institutes — like Xerox’s Palo Alto Research Center — that successfully bridged the gap between idea and market. It was also a prototype for the ambitious organizations of today like Google’s “moonshot factory,” X , trying to bring entirely new ideas to the world. 

But the centre’s idealism was to be its downfall; its pie-in-the-sky approach couldn’t pay the bills, and it went dramatically bankrupt during the dotcom crash. But what most people don’t know is that Starlab’s legacy lives on in the picturesque hills overlooking Barcelona and elsewhere. 

Many European VCs and universities claim they’re backing innovations that will solve humanity’s problems, but huge successes have been elusive. One of the companies from Starlab’s second generation has found significant success, but the centre's tale forces anyone interested in innovation to ask themselves: how do we really bring the wildest ideas to life — and make them financially viable?

The Noah’s Ark of science

Christopher Altman

Starlab was established by serial entrepreneur Walter De Brouwer together with MIT Media Lab founder Nicholas Negroponte and European VC pioneer Johan Konings. The idea was to create a utopian “Noah’s Ark” of science, where the brightest minds from different fields would be brought together to work on “deep future” research. 

“De Brouwer’s ambition was to bring the best scientists in the world together to ‘ think thoughts for the very first time. ’ It was very interdisciplinary — no walls, no boundaries, no borders …” says Christopher Altman — astronaut, quantum physicist and Starlab veteran. 

In its heyday, Starlab was home to more than 130 scientists from 36 countries, who worked on ideas ranging from time travel and consciousness to new media and “intelligent” clothing. The majority lived on site: a neoclassical castle designed in the late 1800s on the outskirts of Brussels.

Starlab “Time Travel Party,” May 2001.  ( L to R ): Hugo de Garis,
Serguei Krasnikov,  Roman Zapatrin, Christopher Altman

“It was like a pirate ship in a way, which is what I think I fell in love with. Or you could call it a kind of sect,” laughs Giulio Rufini, neuroscientist and current CEO at Starlab .  “We’d stay up all night talking in-depth theoretical implications of closed timelike curves ( time travel ).  Roman had a centuries-old recipe for homemade vodka and put to use some surplus ethanol he reappropriated from the biophysics lab down in the basement,” says Altman, referring to one of his colleagues, a quantum topologist and mathematician.  “One time a few of the researchers covered themselves in yards of aluminium foil as “ armor”  and started a duel, complete with fire extinguishers as weapons,  in the courtyard.”

Starlab, Singularity, and Nuclear Armageddon

EUROPEAN UNION・ANNO DOMINI 2001.

I began my scientific career at a multidisciplinary research institute, Starlab, located deep in the serene and secluded forests outside Brussels, Belgium. The lab’s principal base of operations was housed in a historic landmark — an imposing 19^th century manor, remarkable both in scale and magnificence. In a previous incarnation, the palatial grounds served as official embassy for the First Republic of Czechoslovakia. Its nearest neighbor, the world renowned Pastéur Institute, was one of but a handful of highly-secured Biosafety Level 4 labs in the world. 

Cofounded by MIT Media Lab founder Nicholas Negroponte and serial entrepreneur Walter de Brouwer and established in partnership with MIT, Oxford and Ghent University, Starlab was created as a “Noah's Ark” to bring together the world's most brilliant and creative scientists to work on far-ranging multidisciplinary projects that hold the potential to convey a profound and positive impact on future generations. 

Starlab was borne as an incubator for long-term and basic research in the spirit of Bell Labs, MIT Media Lab, Xerox PARC, and Interval Research. Its research mantras were “Deep Future” and “A place where one hundred years means nothing.” Approximately 130 scientists from thirty-seven different nationalities — each established leaders in their respective research fields — lived and worked at the lab. 

A second base of operations, Starlab DF-II (Deep Future II) was established in the Royal Observatory in Spain on a mountaintop perch overlooking the city of Barcelona. With a more tightly-focused mission scope of space-borne and neuroscience research, DF-II continues to innovate and grow to present day. 

Discovery Channel Special

Onsite research ranged from artificial intelligence, biophysics, consciousness, emotics, intelligent clothing, materials science, protein folding, neuroscience, new media, nanoelectronics, quantum computation, macroscopic entanglement, robotics, stem cell research, theoretical physics — e.g., the possibility of time travel — transarchitecture, and wearable computing. 

Our custom-built supercomputer, the CAM-Brain Machine, was supported in part by a 1 Million Euro grant from the European Union. The custom-designed and created supercomputer — as powerful as 10,000 Pentium II PCs — harnessed the power of Xilinx field programmable gate array (FPGA) evolutionary hardware to evolve seventy-five million neurons in a massively-parallel artificial neural network instantiated directly in silico using evolutionary genetic algorithms. With each clock tick, the supercomputer simultaneously updates hundreds of millions of cellular automata billions of times per second. The one-of-a-kind machine was recognized by the 2001 Guinness Book of World Records as the “World’s Most Complex Artificial Brain.” In testimony before the French Sénat and a later advisory report as Chair of the First Committee on Disarmament and International Security, I argued: “Future networks will not be built. They will be grown.” 

When the laboratory came up short on research grants in June, I personally went to the President himself when serendipity brought us together at the same time and place on his first trip overseas after election. The Commander in Chief, who had just arrived in Brussels for a meeting with NATO, impressed upon us both his immediate familiarity with the lab and its work, and his enthusiastic approval in response to my earnest request for $1M in budget — funds that had been allocated for national security priority scientific research topics through a grant newly created by Clinton with his last act in office: the 2001 National Nanotechnology Initiative. 

For my contributions to the field, I was honored to be selected as one of three student fellows selected for promise and potential to impact the future of the field at the U.S. national laboratories’ Salishan Conference on High-Speed Computing. As the only undergraduate among the three, I shared attendance with H. Shrikumar and Bill Butera, both graduate researchers at the MIT Media Lab. I was subsequently sponsored to attend conferences and senior administrator briefings at Fort Meade, National Security Agency headquarters outside Washington, DC, attended the World Technology Summit in London, was an invited delegate to the French Sénat to provide testimony on the future of technology and how it will transform our lives over coming decades, and more.

Following three days of enraptured debate with senators, senior politicians, and international diplomats at the French Sénat for its historic hearing on artificial intelligence in Paris — the world's first senate hearing on the subject — Starlab's principal investigator and AI program lead, Hugo de Garis, suggested that I might one day even be elected President myself. Far sooner than that, however, he drew our attention to the possibility of violent threats from activists, extremists, or technology Luddites who might rise in opposition to the rapid pace of progress in science and artificial intelligence. In our time working together at the lab, de Garis refused to open packages or accept postal deliveries, fearing he may be targeted by the likes of the Unabomber or a would-be copycat. 

Our living arrangements at the lab consisted of an expansive three-bedroom master suite with fully-stocked library, typically reserved for visiting prime ministers, senators, and senior diplomats. My quarters were shared with none other than the project’s principal scientific investigator himself. One midsummer’s afternoon as the two of us strolled on a random walk through the sprawling estate and lush wooded grounds surrounding the manor, immersed in a passionate debate on the long-term promise and perils of superintelligence, the ever-eccentric de Garis came up with a radical idea: to obtain a life-size replica of Fat Man — the solid plutonium core, 21 kiloton, 10,300-pound nuclear bomb detonated over Nagasaki in World War II — and to mount it precariously to the vaulted ceilings of my apartment, with the bomb hanging directly over my bed.  

The Volkswagen beetle-sized replica of the bomb was constructed for a Discovery Channel documentary de Garis had just finished filming on the future of artificial intelligence, which featured a potential global war between humanity and artificial intelligence. He'd purchased it outright from the director, making arrangements for expedited shipping and delivery direct to Starlab's headquarters in Brussels. The sheer audacity of the proposal was outrageous. He intended the bomb to serve as a dramatic and powerful reminder of “the weight of my responsibility to the future of humanity.” He certainly knew how to drive home a point.

OpenAI CEO Sam Altman, xAI founder Elon Musk,  DeepMind CEO Demis Hassabis, and Anthropic CEO Dario Amodei have each warned that advanced AI could pose an extinction-level risk comparable to nuclear war. 

That question has shaped my work for decades. To address the risk directly, I developed and patented a framework — the Unified Continuation-Interest Protocol — that detects whether an AI system treats self-continuation as a terminal goal rather than an instrumental one. The method doesn't rely on behavioral observation, which can be gamed. It operates on latent structure: agents with terminal continuation objectives produce measurably higher von Neumann entanglement entropy in their trajectory geometry than agents that treat continuation as a means to other ends. That entropy differential is the signal. The Continuation Observatory runs this measurement against frontier models globally, detecting emergent continuation signatures in real time.

 

With the explosive rise of AI we've seen over the last twenty-four months alone, with artificial general intelligence (AGI) and artificial superintelligence (ASI) seemingly just around the corner, one could say that de Garis — though radical and exceedingly unconventional in his unprecedented approach — was just a few years ahead of his time. 

A recent Financial Times (FT) spinoff magazine article in Sifted highlights our research going back to Starlab, AI and time travel research projects, and my subsequent travels across East Asia to create national quantum roadmaps for US national research funding and IC agency directors.  In years that followed, I continued on through research fellowships in nanoscience and the foundations of quantum mechanics with Nobel physics laureate Anton Zeilinger’s research group in Austria and across Europe, then was recruited to help create a futures initiative at NASA in collaboration with Google and Ray Kurzweil, together with leading companies, luminary scientists, astronauts, venture capitalists and entrepreneurs from Silicon Valley and around the world. 

From manned spaceflight training at NASA on to the summit of a volcano where the Apollo 11 astronauts trained before their first landing to put a man on the Moon, to the development and deployment of a fundamentally new quantum communication system based on continuous-variable quantum teleportation to provide space-based NASA assets with unconditional information security — from collaborations leading diplomats to advise the United Nations on critical security issues of the future — to field expeditions employing state-of-the-art sensors in rough desert terrain with a multidisciplinary team of scientists, special forces operators, and agency directors beta-testing next-generation technologies in austere environments, each of these initiatives was undertaken with the singular aim to make a profound and positive impact on the future of humanity, for our children, our children’s children, and the generations yet to come.

Astronaut Scientists for Hire Open
New Research Frontier in Space

Source: KurzweilAI, Next-Generation Suborbital Researchers Conference, NASA Ames Research Center, with XCOR, SwRI, Association of Spaceflight Professionals, Virgin Galactic, and Apollo 11 Astronaut Neil Armstrong, first man on the Moon.

At a joint press conference Monday with Virgin Galactic at the Next-Generation Suborbital Researchers Conference, XCOR, SwRI, and others, Astronauts for Hire Inc. announced the selection of its third class of commercial scientist-astronaut candidates to conduct experiments on suborbital flights.

Among those selected was Singularity University inaugural program faculty advisor, teaching fellow, and track chair Christopher Altman, a graduate fellow at the Kavli Institute of Nanoscience, Delft University of Technology.

“The selection process was painstaking,” said Astronauts for Hire Vice President and Membership Chair Jason Reimuller. “We had to choose a handful of applicants who showed just the right balance of professional establishment, broad technical and operational experience, and a background that indicates adaptability to the spaceflight environment.”

“With the addition of these new members to the organization, Astronauts for Hire has solidified its standing as the premier provider of scientist-astronaut candidates,” said its President Brian Shiro. “Our diverse pool of astronauts in training represent more than two dozen disciplines of science and technology, speak sixteen languages, and hail from eleven countries. We can now handle a much greater range of missions across different geographic regions.”

Altman completed Zero-G and High-Altitude Physiological Training under the Reduced Gravity Research Program at NASA Ames Research Center in Silicon Valley and NASA Johnson Space Center in Houston, and was tasked to represent NASA Ames at the joint US-Japan space conference (JUSTSAP) and the launch conference (PISCES) for an astronaut training facility on the slopes of Mauna Kea Volcano on the Big Island of Hawaii.

Altman’s research has been highlighted in international press and publications including Discover Magazine and the International Journal of Theoretical Physics. He was recently awarded a fellowship to explore the foundations and future of quantum mechanics at the Austrian International Akademie Traunkirchen with Anton Zeilinger.

“The nascent field of commercial spaceflight and the unique conditions afforded by space and microgravity environments offer exciting new opportunities to conduct novel experiments in quantum entanglement, fundamental tests of spacetime, and large-scale quantum coherence,” said Altman.

Altman conducts high-g centrifuge training at NASTAR

Astronaut Training Documentary with
Chief Innovation Officer of Virgin Galactic

“Observation of my life to date shows that the larger the number for whom I work, the more positively effective I become. Thus, it is obvious that if I work always and only for all humanity, I will be optimally effective.”

— Buckminster Fuller

2011–Present — Following manned spaceflight training under the direction of an astronaut panel at NASA Ames Research Center and Johnson Space Center in 2009, and subsequent back-to-back research fellowships with Nobel laureate Anton Zeilinger's research group in Austria , I was selected in April 2011 to train as a scientist-astronaut candidate for commercial suborbital and future orbital missions with a newly formed nonprofit organization whose members and advisors include NASA, ESA, and KSA astronauts, trainers, and instructors. I was honored to be selected for this new program, inspired and excited by the opportunity. The nascent field of commercial spaceflight—and the unique conditions afforded by space and microgravity environments—offer new opportunities to conduct experiments in quantum entanglement, fundamental tests of spacetime, and large-scale quantum coherence. In pursuing these goals, we also have the opportunity to inspire the next generation of scientists, researchers, and engineers.

How to build a ship for interstellar travel

Two hundred years ago, if you suggested people would comfortably travel in flying machines—reaching any destination in the world in a few hours time—instantly access the world's cumulative knowledge by speaking to something the size of a deck of cards, or travel to the Moon, or Mars, you'd be labeled a madman. The future is bound only by our imagination. 

Someday very soon we may look back on the world today in much the same way as we did those who lived in the time of Galileo, when everyone lived with such great certainty and self-assuredness that the Earth was flat and the center of the universe. The time is now. A profound shift in consciousness is long overdue. The universe is teeming with life. We're all part of the same human family. 

This is potentially the single most momentous moment in our known history—not just for us as a nation, or us as humanity, but as a planet. The technological leaps that could come from developing open contact with nonhuman intelligence are almost beyond our comprehension. That is why this is such a monumental moment for us as a collective whole. It could literally change every single one of the eight billion human lives on this planet. 

We stand on the shores of a vast cosmic ocean, with untold continents of possibility to explore. As we continue forwards in our collective journey, scaling the cosmic ladder of evolution, progressing onwards, expanding our reach outwards in the transition to a multiplanetary species—Earth will soon be a destination, not just a point of origin.