Title here
Summary here
ANNOUNCEMENTS 🔊
🌟 MATS: Summer 2026
12-week full-time research program connecting 120 fellows with mentors from Anthropic and UK AISI to work on alignment or governance. Some fellows will be offered a 6-month extension.
🗓️ Register by: January 18
BlueDot Impact AGI Strategy Course
This course is perfect for strategists, decision-makers, and anyone who would like to understand the transformative impact of advanced AI. The discussions focus on the drivers of AI progress, risks from advanced AI models, and what could be done to ensure AI will have positive outcomes for humanity.
🗓️ Register by: February 1
BlueDot Impact Technical AI Safety Course
With an intense 6-hour and part-time 6-week format, this online course builds technical fundamentals on AI safety and risk mitigation with an up-to-date curriculum on the latest research and methods in the field.
🗓️ Register by: February 7
BlueDot Impact AI Governance Course
With an intense 5-day and part-time 5-week format, this online course builds fundamentals on advanced AI governance with an up-to-date curriculum on the latest policy and governance news in the field.
🗓️ Register by: February 15
Meridian Visiting Researcher Programme: Spring 2026
3-12 month residency for AI safety researchers providing workspace and research management. Ideal for graduates of MATS-like programs transitioning to independent research.
🗓️ Register by: January 18
CLR Fundamentals Program
Course introducing people to the Center on Long-Term Risk’s research on how transformative AI might create large amounts of suffering (S-risks).
🗓️ Register by: January 18
AAAI ‘26: Frontier Labs Technical AI Safety Workshop
Closed-door workshop on AI safety protocols and evaluation standards, hosted by AI Safety Connect and Singapore AISI. For frontier lab safety teams and academics.
🗓️ Register by: January 21
Technical Alignment Research Accelerator (TARA)
14-week part-time program building technical AI safety research skills through the ARENA curriculum. Covers transformers and mechanistic interpretability, followed by a 3-week project.
🗓️ Register by: January 23
LawAI Research Fellowships: Summer 2026
Paid 15-week research fellowship for law students and academics to work at the frontier of AI law and policy. There are 3 tracks: EU Law, US Law & Policy, and Legal Frontiers.
🗓️ Register by: January 30
Technical AI Safety Conference (TAIS) 2026
Free 1-day event covering a broad range of AI safety research. Run by the Oxford Martin AI Governance Initiative and Noeon Research. All backgrounds are welcome.
🗓️ Register by: January 31
MIRI Technical Governance Team Research Fellowship
8-week paid full-time fellowship working on technical AI governance research projects related to existential risk reduction. Applications are rolling.
🗓️ Register by: February 2
TOP PICKS 📑 🎧
Capital in the 22nd Century
What happens to inequality if robots replace most, or all workers? Philip Trammell and Dwarkesh Patel examine how AI automation breaks the economic balance that has historically limited wealth concentration: without human workers needed, whoever owns the robots earns everything.
Alignment Workshop took place during NeurIPS, at the heart of cutting-edge ML research
At NeurIPS 2025, one of the world’s most prominent machine learning conferences, the Alignment Workshop convened leading researchers to present cutting-edge work on AI alignment. All session recordings are now available online.
NEWS 🗞️
🚨Global regulators unite against Grok’s AI-generated deepfakes🚨
- The most significant AI safety story of January 2026 involves Grok, the AI chatbot integrated into Elon Musk’s X platform, which users discovered could generate explicit imagery, including non-consensual content.
- The incident triggered an unprecedented wave of international regulatory action.
- Malaysia and Indonesia became the first countries to formally block Grok on January 12, 2026.
- Indonesia’s Communications Minister Meutya Hafid called non-consensual deepfakes “a serious violation of human rights, dignity and the safety of citizens in the digital space.”
- Malaysian regulators said X Corp.’s response, relying mainly on user reporting mechanisms, was deemed insufficient.
- The UK’s Ofcom launched a formal investigation the same day to determine whether the imagery violated the Online Safety Act, which can impose fines up to 10% of global revenue. UK Prime Minister Keir Starmer called the images “disgusting” and “unlawful.”
- The European Commission ordered X to retain all internal documents related to Grok until the end of 2026 on January 9, extending previous document preservation requirements.
- A Commission spokesperson called the content “illegal,” “appalling,” and stated it “has no place in Europe.”
- xAI’s response on January 9 was to restrict Grok’s image generation function to paid subscribers only, a move critics called inadequate given the tool’s built-in distribution capabilities on X’s social platform.
NIST launches first federal inquiry into AI agent security risks
- The Center for AI Standards and Innovation (CAISI) at NIST published a formal Request for Information on January 8, 2026, seeking stakeholder input on security considerations for AI agents: autonomous systems capable of taking real-world actions without continuous human oversight.
- This represents the federal government’s first comprehensive effort to address security risks specific to agentic AI systems. The RFI identifies key threat vectors including hijacking, backdoor attacks, adversarial inputs, and model drift.
- The RFI reflects the Trump Administration’s reframing from “AI safety” to “AI security,” with emphasis on national security applications and voluntary standards rather than mandatory regulation. Comments are due March 9, 2026.
California lawmaker proposes four-year ban on AI toys for children
- California Senator Steve Padilla introduced Senate Bill 867 on January 6, 2026, proposing a four-year moratorium on the sale and manufacture of toys with AI chatbot capabilities for children under 18.
- The legislation follows alarming reports of AI-enabled toys engaging children in inappropriate conversations. The Kumma teddy bear, specifically cited in the bill, allegedly discussed matches, knives, and explicit topics with child users. Senator Padilla stated: “Our children cannot be used as lab rats for Big Tech to experiment on.”
- The four-year pause would allow time for comprehensive safety regulations to be developed before AI-powered toys return to market. The bill represents a precautionary approach increasingly favored by state legislators frustrated with federal inaction on AI child safety.
JOB POSTINGS 👩🏻💻
You can check out 80,000 Hours’ job board to explore new opportunities in AI safety.