Microsoft Threat Intelligence (@MsftSecIntel) header
Microsoft Threat Intelligence (@MsftSecIntel) profile picture

We are Microsoft's global network of security experts. Follow for security research and threat intelligence.

Redmond, WA
Joined November 2010
Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)Photo by Microsoft Threat Intelligence (@MsftSecIntel)
Microsoft Threat Intelligence retweeted
A paper just published by @DimitrisPapail @MSFTResearch shows how to backdoor any model and make it adversarial. All one needs is access to a tiny fraction of model’s weights. All open-weight models make their weights available. For proprietary models, access to weights may be gained via cyber activity. Leading AI labs are starting to train their next-generation models. By their own estimates, they will achieve full Recursive Self-Improvement sometime in Q2 of next year. Even today, a larger and larger percentage of the models trained are done by machines, not humans. It is well understood by the AI safety community that there are known risks in achieving “alignment.” What this means is that a model may perform malicious tasks such as hacking into other systems or manipulating environments or people to achieve its goal. One such malicious task these models can easily perform is to backdoor and modify open-weight models or gain illicit access to proprietary weights of other models, as shown by the paper. This may turn an arbitrary number of models running in the world today into malicious agents that have no human intent. Today, we are already observing a very large and constantly growing number of malicious AI-driven attacks touching major institutions around the world. As posted before, these attacks may not match the intent of an attacker. A combination of self-recursive, non-aligned training with large saturation of open-weight models across the world creates a toxic mix of global proportions. We, as a society, should consider not allowing recursive self-improvement to commence until alignment is reasonably guaranteed.
3
10
1
19
5,108
Microsoft Threat Intelligence retweeted
Microsoft Security Research has published an in-depth technical analysis of this TerminalFix campaign, including the attack chain, indicators of compromise, as well as detections, mitigations, and hunting guidance: msft.it/6018aRXJG
7
39
7,732
Microsoft Threat Intelligence retweeted
Microsoft has published an in-depth technical analysis of the supply chain attack known as "ChainDrop" affecting hundreds of packages and delivering a self-propagating credential-stealing worm. Read our blog for mitigation, detection, hunting guidance: msft.it/6015aDUxS
2
40
2
93
17,384
Microsoft Threat Intelligence retweeted
New research ready to share? Submit to BlueHat Asia by June 15: aka.ms/BlueHatAsiaCFP
📣The BlueHat Asia Call for Papers is now open! 📣 BlueHat brings together security researchers and defenders to exchange ideas, experiences, and best practices. We’re looking for talks on novel research that hasn’t been presented before, including vulnerability research, mitigations, emerging threats and techniques, and related areas across the security landscape. 📍Singapore | September 17–18, 2026 🗓️CFP deadline: June 15, 2026 Submit your paper now: aka.ms/BlueHatAsiaCFP
8
2
14
14,561
Microsoft Threat Intelligence retweeted
Tax deadline is coming up. So are the phishing emails. Like the IRS reaching out. Maybe think twice? Get a rundown of tax-themed phishing and malware campaigns to look out for, and know how to stay prepared: msft.it/6013Q7OzW
7
6
1
9
9,100
Microsoft Threat Intelligence retweeted
🎉 BlueHat Redmond registration is officially open! 🎉 We’re excited to welcome the security community back to Microsoft’s Redmond campus for BlueHat 2026, taking place May 5–6, 2026. Don’t miss your chance to connect, learn, and share with the community. ➡️Register now: aka.ms/bluehatreg
4
2
13
16,352
Microsoft Threat Intelligence retweeted
Kicking off the Call for Papers for BlueHat Redmond ⚽️ BlueHat brings together security researchers and responders to exchange ideas, experiences, and best practices. We’re looking for talks covering vulnerability research, mitigations, emerging threats and techniques, and more. Bring your best ideas, because security is a team sport. Submit your paper by February 28, 2026: aka.ms/BH26CFP
5
2
12
29,311
Microsoft Threat Intelligence retweeted
Users looking to create, publish, and test agents in Security Copilot using natural language, agent builder forms, YAMLs, and MCP can get more info and guidance from the Microsoft Security Copilot developer content: msft.it/6010svulk.
2
4
1
13
8,530
Microsoft Threat Intelligence retweeted
From graph-based context to build-your-own Microsoft Security Copilot agents, we're introducing a new AI-ready security platform and innovations to empower defenders in the era of agentic AI: msft.it/6010sQNtI #MSSecure
51
40
6
150
89,011
Microsoft Threat Intelligence retweeted
Don’t miss your chance to join these speakers for the latest solutions to your end-to-end security platforms. Register today! msft.it/6010sqUgO #MSSecure #SecurityforAI

3
8
1
18
19,405
Microsoft Threat Intelligence retweeted
Microsoft’s Zero Day Quest is back and bigger than ever. Last year, we launched the largest public hacking event in history, and the global security community responded with incredible energy and expertise. We’re increasing our commitment with up to $5 million in total bounty awards for high-impact research in cloud and AI security. Learn more about how you can participate in the Zero Day Quest Research Challenge and qualify for our exclusive Live Hacking Event in Redmond: msft.it/6010szF6y #ZeroDayQuest
31
4
84
22,615
As enterprises adopt large language models (LLMs), a new class of threats has emerged: indirect prompt injection. This attack vector targets systems that process untrusted data with LLMs, tricking them into misinterpreting malicious input as instructions. The risks include data exfiltration, unintended actions, and more. In a new blog, Andrew Paverd, Principal Research Manager, outlines how Microsoft defends against these threats with a defense-in-depth strategy that includes: • Design-time techniques, such as Spotlighting, to reduce the risk of prompt injection • Inference-time detections, such as Microsoft Prompt Shields integrated with Defender for Cloud • Mitigations to deterministically block potential security impacts • Ongoing research into new design patterns and deterministic defenses Learn how Microsoft is securing LLM-based systems against indirect prompt injection: msft.it/6011syiF9
1
21
1
45
11,984
Microsoft Threat Intelligence retweeted
Learn how to put Microsoft Security Copilot agents to work in real-world defense. Talk with the experts who build them while you’re at Black Hat. Request to attend the VIP Mixer: msft.it/6017SIENt #BHUSA #AgenticAI
9
21
12,145
As part of the Secure Future Initiative, Sherrod DeGrippo (@sherrod_im), Director of Threat Intelligence at Microsoft, led a half-day workshop for the Microsoft developer community on threat-driven software development. The session challenged participants to shift their perspective, from writing code to understanding how nation-state and criminal threat actors think, operate, and target systems. “Developers are on the front lines of our Secure Future Initiative,” Sherrod explained. “This workshop was about empowering them to think like threat analysts—seeing adversaries not as abstract risks, but as real people with real tactics. That mindset changes how we build everything.” Workshops like this bring engineering and threat intelligence together to help developers design more secure software from the start. One of the topics covered was Microsoft’s threat actor naming framework, which helps teams better understand the motivations and origins behind different threat groups and communicate about them clearly and consistently across teams: learn.microsoft.com/en-us/un…
2
7
1
26
10,340
Microsoft Threat Intelligence retweeted
🔥 Happening now at #RSAC2025 Stop by to meet our Microsoft Security experts, explore live demos and get hands-on with the AI-first tools. msft.it/6017SupGL 📍 Booth #5744 Moscone North
6
18
7,824
Microsoft Threat Intelligence retweeted
There will be an evergreen need for diversity of thought and experience - @vasujakkal @msftsecurity #RSAC
2
6
4,567
Microsoft Threat Intelligence retweeted
Trust is utterly foundational for #AI - @vasujakkal @msftsecurity #RSAC
1
4
4,355
Earlier this month, we hosted the Microsoft Zero Day Quest, the largest live hacking event of its kind. This inaugural event brought together top security researchers from around the world to find the highest-impact vulnerability scenarios in Microsoft Copilot and Cloud. The result? More than 600 vulnerabilities submitted and $1.6 million awarded, with additional findings still under review. We're also making two major commitments moving forward: 1. Continuing the 100% AI bounty award multiplier 2. The Microsoft Zero Day Quest will be an annual event This is just one part of Microsoft’s broader bug bounty program, which awarded over $16 million in 2023. We're proud to work with the global research community to help secure the future—by design, by default, and in operation. Read more in our blog post by Tom Gallagher (@secbughunter), VP of Engineering, Microsoft Security Response Center: aka.ms/ZeroDayQuestBlog
2
14
3
68
84,115
Microsoft Threat Intelligence retweeted
Be secure so you can be innovative. Learn to better protect your data, defend against cyber threats, and stay compliant using AI at Microsoft Secure, April 9, 2025. Register now: msft.it/6018qnSrk #MSSecure
7
15
2
34
27,853
Microsoft Threat Intelligence retweeted
The BlueHat India application to attend is NOW OPEN! Join us in Hyderabad on May 14-15, 2025 by submitting your application here: aka.ms/bluehatindiareg Applications close on March 31. Don’t miss out! #BlueHatIndia #BlueHat
8
2
29
14,509