Andrew W Neal

Apart from the silly rightwing quips by the author, this is a fascinating strategic analysis.

'Putin is losing the war, so prepare for escalation' share.google/0UCU79em6QFx...

Hear me now! Well, soon...

Going to Salzburg to talk about Large Language Model research in IR was fun!

Analysis is based on sentence-level semantic search of inductively-discovered threat topic groups, filtered to include the word 'threat'. Data sourced from the Edinburgh National Security and Defence Documents dataset. datashare.ed.ac.uk/handle/10283...

State-sponsored threats:
Both espionage and sabotage emerge as prominent concerns in 2025—espionage doubling from the 2020-2023 baseline whilst sabotage appears as an entirely new threat category—reflecting intensified state competition and the targeting of critical national infrastructure.

Missile threat salience rose 90% in 2025, likely driven by proliferation of advanced ballistic and hypersonic capabilities amongst adversarial states and the erosion of traditional arms control frameworks.

Hybrid warfare salience increased 67% by 2025, indicating growing governmental concern over coordinated grey-zone operations that blur the boundaries between peace and conflict through disinformation, cyberattacks, and economic coercion.

AI and emerging technologies:
Artificial intelligence appears as a discrete security threat for the first time in 2025, marking official recognition of AI's dual-use potential and autonomous systems vulnerabilities in an increasingly technology-dependent security landscape.

Military/Traditional Threats:
Military threats peaked early (2002-2003) and have maintained consistent but lower-level attention, with recent uptick in 2025, reflecting renewed conventional security concerns.

Climate: emerging priority
Climate-related security concerns were absent from early documents but emerged in 2009 and peaked in 2020. Notably, explicit mention climate threats have disappeared entirely from 2025 documents, suggesting a possible shift in framing or priorities.

Nuclear: persistent concern
Nuclear threats have maintained steady relevance throughout the period, with notable spikes in 2008 and consistent presence in recent documents.

Cyber: rise and plateau
Cyber threats first gained significant attention in 2009 and peaked in 2015. Unlike terrorism, cyber threats have remained consistently present but at lower levels recently.

Terrorism: declining
Terrorism emerged as the primary security concern post-9/11, peaking during 2008-2009 and maintaining high levels through to 2018. However, there's been a marked decline in recent years.

In response to the new UK National Security Strategy and SDR, here's my analysis of changing threat topics in UK national security and defence documents since 1998.
@uoe-sps.bsky.social @scga.bsky.social @rethinkingsecurity.bsky.social @chathamhouse.org @rusi.bsky.social @edinburghpir.bsky.social

Securitization, Politicization and Social Differentiation in History

My chapter "NATO accession, democratization, and security sector reform" is out! I examine SSR & democratisation in CEE, highlighting how historical positionality shapes understanding.

Explore the book (ed. @tbonacker.bsky.social): edin.ac/3FOOB4c
#SecurityStudies #Democratisation #PoliticalScience

Happy to win this little internal prize for my National Security and Defence Documents Dataset (1987-2024) 😊

datashare.ed.ac.uk/handle/10283...

🤣

'You could go off to the Shetlands with a shovel and cut off the internet to the Faroe Islands fairly easily if you wanted to'

I somehow managed to say this BBC Radio Scotland. I hope nobody gets any ideas 😂

They don't call me Professor of International Security for nothing you know.

Lord Robertson, former head of NATO, was a highlight, with his Scottish raconteur style. Also my first time seeing a PM up close. Pro-Palestinian protesters held a diligent presence outside, good for them. I spoke on a great mixed panel on energy security.

A fascinating few days at the London Defence Conference. Quite an eye opening range of views from a mixed defence industry/policy/academic crowd, very hawkish at times (perhaps unsurprisingly). I had to remind myself that it's just one sector among many.

OMG. The revolution may not be televised, but the war will be memed on TikTok.

Introducing the National Security and Defence Documents Dataset 1987-2024 A new dataset aims to compile and make available for analysis every national security and defence policy document ever published anywhere in the world. The dataset’s creator, Andrew W. Neal, explai…

An updated dataset aims to make available for analysis every national security & defence policy document ever published anywhere globally. @andrewwneal.bsky.social explains how the collection reveals trends in public security discourse & threat perceptions.

rethinkingsecurity.org.uk/2024/11/14/n...

A PhD student was snatched by masked officers in broad daylight. Then she was flown 1,500 miles away | CNN Tufts University PhD student Rumeysa Ozturk was walking alone Tuesday night to meet friends at a dinner where they would break their 13-hour fast when six plainclothes officers suddenly encircled her ...

Truly terrifying: arbitrary, vindictive, discrimatory, cruel, callous, and hypocritical. Not to a mention a compete disregard for the law or courts.

www.cnn.com/2025/03/29/u...

I have tariffed
the penguins
that are on
Heard Island

and which
you were probably
assuming
did not export goods

forgive me
they were taking advantage of us
so cunning
and so cold

@dandrezner.bsky.social not mincing his words

@rethinkingsecurity.bsky.social @scga.bsky.social @mybisa.bsky.social @isanet.bsky.social @uoe-sps.bsky.social @chathamhouse.org @rusi.bsky.social @britishacademy.bsky.social @peacerep.bsky.social @duckofminerva.bsky.social @warontherocks.bsky.social @nupinytt.bsky.social @sipri.org

Download now: Edinburgh DataShare doi.org/10.7488/ds/7...

Please share!

#InternationalRelations #DefencePolicy #SecurityStudies #DataScience #ComparativePolitics #NLP

- No advanced ML knowledge required to use the basic search features
- Ideal for researchers in international relations, security studies, defence policy, as well as policymakers, civil society organisations, and educational institutions interested in global security priorities.

Technical skills needed:
- Basic Python knowledge for running the semantic search tools
- Familiarity with Jupyter notebooks to explore the data
- Understanding of NLP concepts helpful but not essential

- Natural Language Processing tools for analysing document content
- Jupyter notebooks for data exploration and visualisation
- Reproducible code for computational text analysis