@Defense_AIOi
iAccount based inSwitzerland
About this account
- Account based in
- Switzerland
- Connected via
- Switzerland App Store
Account-level information from X, not a live location or the device used for a specific post.
The Defense AI Observatory (DAIO) monitors and analyses the use of artificial intelligence by armed forces | Funded by @dtecbw I Operated by @heikoborchert
Hamburg
Joined January 2021
- Tweets791
- Following217
- Followers1.6K
- Likes478
Pinned Tweet
📢+++ OUT NOW: THE VERY LONG GAME+++📚
...3...2...1: We are truly excited to inform you that "The Very Long Game: 25 Studies on the Global State of #DefenseAI" has been released today. 🧵 with links to all chapters below
Get your #OpenAccess copy here: link.springer.com/book/10.10…
📣Trusting Allied Machines📚
In our latest paper, LTC Sanghyun Ahn, ROK Army, argues that allied nations in the APAC region can build a regional assurance regime for defense AI "without wating for the hegemon to supply it." For more, see: defenseai.eu/daio_study2637_…
📣 Reimagining Horizons: Defense AI in the Philippines🇵🇭The Philippines, writes Keanu Paul Sygaco in our latest study, "must pair reactive measures against immediate threats with proactive efforts" to advance defense AI. For more, see his study: defenseai.eu/daio_study2638_…
MOSA is about much more than only creating open interfaces, writes @HeikoBorchert. For more, see👇🏻👇🏻
MOSA: Es geht um viel mehr als offene Schnittstellen
Um anpassungsfähiger zu werden, setzen Streitkräfte verstärkt darauf, monolithische Produkte durch einen modularen und offenen Systemansatz (#MOSA) zu ersetzen. Das Bundesministerium der Verteidigung hat hierzu kürzlich seine eigene Weisung erlassen.
Offene Schnittstellen sind für MOSA zwar wichtig, doch es geht, wie ich im heutigen Standpunkt für den @Tagesspiegel Background Sicherheit & Verteidigung darlege, um deutlich mehr, denn MOSA
1️⃣ greift tiefer und bedeutet, die Eins-zu-Eins Verbindungen in bestehenden Lösungen durch Module zu ersetzen, die in Verbünden kombiniert und durch dynamisch anpassbare Abläufe gesteuert werden;
2️⃣ erfordert die anpassungsfähige Koordination eben dieser Module, die lagespezifisch aufeinander abgestimmt werden können, um den Gegner zu überraschen und ihm eine gewünschte Vorgehensweise (Course of Action, CoA) aufzuzwingen;
3️⃣ bedingt weitreichende rüstungsindustrielle Anpassungen, denn die Idee, militärische Fähigkeiten wie Mosaikbausteine verteilt zu organisieren, ist bislang nur unzureichend abgebildet;
4️⃣ ist als Designprinzip auch für die umfassende Sicherheit von zentraler Bedeutung, um z.B. Aufklärungsinformation für die Drohnenabwehr allen staatlichen und nicht-staatlichen Sicherheitspartnern zur Verfügung zu stellen.
Link zum Beitrag: background.tagesspiegel.de/b…
cc @svenweizenegger @MW42522637 @Ce_Moll @_schuetzt @Konflikt_Sicher @Bjoern__M
Turning Experimentation Into Military Advantage
Portugal is following a "distinctive model among NATO and EU nations" that emphasizes the role of experimentation in adopting defense AI, writes MajGen Joao Vicente, Portuguese Air Force.
For more, see defenseai.eu/daio_study2636_…
Defense AI at the 2026 #AISummitBudapest 👇🏻👇🏻
"When AI tools are becoming collaborative partners."
This is the leitmotif of the 2026 #AISummitBudapest, which will take place on 7-8 September. Organized for the fourth time, the conference has turned into Central Europe's premier platform to discuss all things AI....
This year the conference also runs a dedicated workstream on AI in space and defense. Together with Gergely Németh, CEO of KIVI, and MajGen Laszlo Kovacs, HUN MoD, we will explore
🤖 how armed forces are using AI,
🤖 to what extent current use cases describe evolutionary or revolutionary pathways to using AI, and
🤖 what challenges armed forces face when cooperating with non-traditional defense partners.
If you are in Budapest on 7-8 September, please join us on Monday, 7 Sept, 1430 (Quantum Stage). We look forward to welcoming you.
For more on the program and to register, see: aisummitbudapest.hu/
Interesting in view of our 2021 study Beware the Hype
defenseai.eu/daio_beware_the…
“Russia’s war in Ukraine and modern warfare” is a great companion to “Preparing for war” by @Olivier1Schmitt. I particularly like the versatility model developed by @mikaelweissmann and Niklas Nilsson (see pic)
Use the link in Mikael’s original post below 👇🏻 to get your copy of this open-access volume
cc @Mauro_Gilli @aa_gilli @MichaelRaska @CyberMyri @ZaccagniniIvan
Defense AI Observatory retweeted
"Ambivalent Acceleration" is now also available on @SSRN. Get your copy at papers.ssrn.com/sol3/papers.….
Need an #AI refresher?
As an early mover in #AIGovernance, SIPRI has you covered. Explore these recent publications and commentaries to get back up to speed:
🟣 SIPRI Yearbook 2026: Artificial Intelligence and International Peace and Security
🔗sipri.org/yearbook/2026
🟣 Mapping the Military AI Industry
🔗 bit.ly/4atY7po
🟣 Responsible Procurement of Military Artificial Intelligence
🔗bit.ly/4x8jV2X
🟣 Autonomous Weapon Systems and AI-enabled Decision Support Systems in Military Targeting: A Comparison and Recommended Policy Responses
🔗 doi.org/10.55163/YQBY3151
🟣 Towards Multilateral Policy on Autonomous Weapon Systems
🔗 doi.org/10.55163/VLLZ7787
🟣 Addressing the Risks that Civilian AI poses to International Peace and Security: The Role of Responsible Innovation
🔗 doi.org/10.55163/OXHH2798
🟣 Impact of Military Artificial Intelligence on Nuclear Escalation Risk
🔗 doi.org/10.55163/FZIW8544
🟣 Bias in Military Artificial Intelligence and Compliance with International Humanitarian Law
🔗 doi.org/10.55163/NLWV5347
🟣 Commentary: Before it’s too late: Why a world of interacting AI agents demands new safeguards
🔗 bit.ly/3OzMkh8
🟣 Commentary: Responsible behaviour in military AI starts with responsible procurement
🔗 bit.ly/4qyyCJm
#ArmamentAndDisarmamentWeek @TMCAsser #ArtificialIntelligence #AIGovernance @GCREAIM @UN_Disarmament @Defense_AIO @BoulaninSIPRI
📢 Small Force, Big Data: Defense AI in New Zealand📚
New Zealand's defense #AI challenge is wheter the country can "build enough competence to remain useful, trusted, and sovereign inside allied systems" writes @ReubenSteff in his latest study
See: defenseai.eu/daio_study2635_…
Stress-testing defense assumptions with the defense metaverse and powerful #redAI tactics. We are happy to share the latest piece by @HeikoBorchert for @HorizonDefence
Are You REDy? Using Red AI Tactics and the Defense Metaverse to Augment Force Design
"The European Central Bank is stress-testing the resilience of the European financial system against hypothetical financial and economic shocks. The Defence Metaverse and red AI tactics provide a novel military equivalent to stress-test national and European defence assumptions, to develop concepts and capabilities for the new age of protracted warfare characterised by adversaries deliberately depleting European military capabilities."
This is the key argument of my latest paper for @HorizonDefence.
For more, see tdhj.org/blog/post/red-ai-ta….
Many thanks @MW42522637 @Pastey_n and the @HorizonDefence for your help.
@KubiliusA @Mauro_Gilli @MichaelRaska @Konflikt_Sicher @Bjoern__M @VorgBeob
Supporting capability development and future force design with the defense metaverse and AI tactics. Listen in to the discussion with @bernhardkast and @HeikoBorchert on @MilHiVisualized 👇🏻
1️⃣ You are about to launch a new C-UxS solution but you don't know if better sensors, more advanced effectors, or a more sophisticated combination will make the difference
2️⃣ Your fleet is about to embark on a new convoy protection mission but you are worried about optimal weapon load given limited magazine depth
3️⃣ You face growing parliamentary scrutiny over raising defense spending and want to make the case that AI can help augment legacy systems but you don't know how to optimally combine them with new systems
If these are the questions you are dealing with, consider exlporing the use of the defense metaverse, a high-fidelity virtual twin of the battlefield, and AI tactics as new instruments to advance capablitiy development and future force design and support procurement decisions.
For more see the interview with @bernhardkast on his YouTube channel @MilHiVisualized
youtube.com/watch?v=YesuhCQK…
In “Ambivalent Acceleration” @HeikoBorchert provides the latest comparative overview of our ongoing analysis of defense AI around the world - now covering 37 countries
AI Goes to War and Is Not (Yet) Winning
This is the core message of "Ambivalent Acceleration," my new comparative assessment of the current state of defense AI in (now) 37 countries. Frank Schmiedchen et al. will publish it in their forthcoming volume "Living with AI: State, Applications, and Challenges of Artificial Intelligence," which should be available from Springer later this year. This chapter updates the findings of all 25 case studies published in "The Very Long Game," and adds new insights from the latest DAIO case studies on PAK, POL, IDN, ROU, BEL, BRA, and all six GCC nations
Contrary to the prevailing rhetoric, AI is still primarily considered an extension of existing military systems and platforms, thus prolonging the imaginaries (see yesterday's post "Goodbye California – Hello Kyiv"), concepts, and practices developed to use them.
Yes, we are witnessing an accelerated push for defense AI driven by war(s), battlefield deployment, and increasing defense spending and private defense funding. But the outcome is ambivalent:
First, emulation is even more dominant than two years ago, albeit with a different twist as the fear of missing out on how Ukraine is using defense AI (FOMO-U) is creating a new role model. And so is, among others, the 5-day war between PAK and IND, as I argue in the paper.
Second, developing AI against the background of existing systems and platforms shapes AI in line with their prime tasks. This is strengthening "defense verticals." That's why software is not "eating" defense hardware, as the paper discusses.
Finally, concerns about digital sovereignty are significantly more pronounced today than two years ago. This creates a geostrategic dilemma as the partner you might want to work with for geopolitical reasons is the one that limits your leeway for geoeconomic reasons.
And now…..enjoy reading the paper. borchert.ch/wp-content/uploa…. Thank you
cc @KarasikTheodore @Lauren_A_Kahn @mchorowitz @paul_scharre @KofmanMichael @HoansSolo @Mauro_Gilli @aa_gilli @Konflikt_Sicher @HorizonDefence @Ce_Moll @breitenbauch @sambendett @Adib_Moghaddam @jessicadorsey @WarintheFuture
Defense AI Observatory retweeted
We're proud to have co-organised #TMAF2026 alongside outstanding partners, bringing together military professionals, academics, innovators, and future leaders to discuss the future of defence and military education.
A huge thank you to all speakers, authors, participants, and supporters who made this year's forum a success.
What do you see as the biggest challenge—or opportunity—for defence and military education in the next decade?
We're already looking ahead to #TMAF27. 🚀
#Defence #MilitaryEducation #SecurityStudies #Innovation #Leadership #FutureOfDefence #TDHJ @thermilak @Pastey_n
@DrBrittainHale @Bernd_Schulyok
@HeikoBorchert @ALauner1 @lukas_bittner @ETH_en @Bundesheerbauer
youtu.be/_DshNzXvINQ?si=FDbg…
Defense AI Observatory retweeted
Ongoing debates about technology's impact on warfighting easily overlook one key aspect: #skills #proliferation. Violent non-state actors are among the first movers or, as Carl Peterson writes in his latest piece for @MilitantWire (link below):
"In this hothouse of drone innovation, both technological and tactical, criminal and rebel groups have seen the opportunity to increase their capability by either sending existing mebmers to serve in the Ukranian or Russian militaries, recuriting returned fighters into their organizations, or otherwise to purchase the services of returned fighters on a mercenary basis. Currently, criminal organizations are pursuing all three pahts of bringing skiled drone operators into their organizations...."
And now think about current discussions on civil preparedness and critical infrastructure protection....
militantwire.com/p/the-drone…
Defense AI Observatory retweeted
Totally agree. If you always want to copy/imitate, the competitive advantage is to your disadvantage unless the industry has growing marginal costs — which is not true for software. For europe, this is an opportunity but also a challenge because to devise the next thing you need R&T and financial and cultural risk-acceptance.
The point is not to finde a #European clone of #Palantir. Europe must think differently: "not chasing after those who believe they can decide faster, but developoing solutions that produce better decisions."
Thank you @Lorenzomonfreg, @ilfoglio_it. Links below
The point is not to finde a #European clone of #Palantir. Europe must think differently: "not chasing after those who believe they can decide faster, but developoing solutions that produce better decisions."
Thank you @Lorenzomonfreg, @ilfoglio_it. Links below