Luca Righetti

Kamilė Lukošiūtė, John Halstead, Luca Righetti

AI companies and governments are increasingly concerned about frontier AI systems enabling cybercrime, yet defining meaningful capability thresholds requires knowing the scale of cybercrime today. Current estimates of global cybercrime damages vary from tens of billions to tens of trillions of dollars, with little systematic evaluation of their reliability. We establish a more rigorous baseline by surveying 27 existing estimates, critically evaluating their methodologies, and constructing a composite estimate from three independent sources: a nationally representative UK business victimization survey scaled globally, US individual victimization data scaled globally, and global cybersecurity spending figures. Large-sample victimization surveys capture losses directly from victims, avoiding both the reporting bias in law enforcement and industry databases and the heavy modeling assumptions of macroeconomic approaches. We focus on quantifiable economic damages -- direct losses, response costs, and defense spending -- excluding harder-to-measure costs such as intellectual property theft and reputational damage. We estimate total global cybercrime damages at approximately \$500 billion USD annually (90% CI: \$100 billion-\$1 trillion). At this baseline, an AI-driven increase of about 20% would add \$100 billion or more, reaching thresholds some companies identify as warranting additional mitigations. However, cybercrime data remains too incomplete for such incremental increases to be directly detectable, and defensive applications of AI may partially offset offensive gains. Our methodology narrows plausible damage estimates from multiple orders of magnitude to a more confident baseline, providing a foundation for decision-making about AI-related cybercrime risk.

Shen Zhou Hong, Alex Kleinman, Alyssa Mathiowetz et al.

Large language models (LLMs) perform strongly on biological benchmarks, raising concerns that they may help novice actors acquire dual-use laboratory skills. Yet, whether this translates to improved human performance in the physical laboratory remains unclear. To address this, we conducted a pre-registered, investigator-blinded, randomized controlled trial (June-August 2025; n = 153) evaluating whether LLMs improve novice performance in tasks that collectively model a viral reverse genetics workflow. We observed no significant difference in the primary endpoint of workflow completion (5.2% LLM vs. 6.6% Internet; P = 0.759), nor in the success rate of individual tasks. However, the LLM arm had numerically higher success rates in four of the five tasks, most notably for the cell culture task (68.8% LLM vs. 55.3% Internet; P = 0.059). Post-hoc Bayesian modeling of pooled data estimates an approximate 1.4-fold increase (95% CrI 0.74-2.62) in success for a "typical" reverse genetics task under LLM assistance. Ordinal regression modelling suggests that participants in the LLM arm were more likely to progress through intermediate steps across all tasks (posterior probability of a positive effect: 81%-96%). Overall, mid-2025 LLMs did not substantially increase novice completion of complex laboratory procedures but were associated with a modest performance benefit. These results reveal a gap between in silico benchmarks and real-world utility, underscoring the need for physical-world validation of AI biosecurity assessments as model capabilities and user proficiency evolve.

Tegan McCaslin, Jide Alaga, Samira Nedungadi et al.

Evaluations of dangerous AI capabilities are important for managing catastrophic risks. Public transparency into these evaluations - including what they test, how they are conducted, and how their results inform decisions - is crucial for building trust in AI development. We propose STREAM (A Standard for Transparently Reporting Evaluations in AI Model Reports), a standard to improve how model reports disclose evaluation results, initially focusing on chemical and biological (ChemBio) benchmarks. Developed in consultation with 23 experts across government, civil society, academia, and frontier AI companies, this standard is designed to (1) be a practical resource to help AI developers present evaluation results more clearly, and (2) help third parties identify whether model reports provide sufficient detail to assess the rigor of the ChemBio evaluations. We concretely demonstrate our proposed best practices with "gold standard" examples, and also provide a three-page reporting template to enable AI developers to implement our recommendations more easily.

Owain Evans, Owen Cotton-Barratt, Lukas Finnveden et al.

In many contexts, lying -- the use of verbal falsehoods to deceive -- is harmful. While lying has traditionally been a human affair, AI systems that make sophisticated verbal statements are becoming increasingly prevalent. This raises the question of how we should limit the harm caused by AI "lies" (i.e. falsehoods that are actively selected for). Human truthfulness is governed by social norms and by laws (against defamation, perjury, and fraud). Differences between AI and humans present an opportunity to have more precise standards of truthfulness for AI, and to have these standards rise over time. This could provide significant benefits to public epistemics and the economy, and mitigate risks of worst-case AI futures. Establishing norms or laws of AI truthfulness will require significant work to: (1) identify clear truthfulness standards; (2) create institutions that can judge adherence to those standards; and (3) develop AI systems that are robustly truthful. Our initial proposals for these areas include: (1) a standard of avoiding "negligent falsehoods" (a generalisation of lies that is easier to assess); (2) institutions to evaluate AI systems before and after real-world deployment; and (3) explicitly training AI systems to be truthful via curated datasets and human interaction. A concerning possibility is that evaluation mechanisms for eventual truthfulness standards could be captured by political interests, leading to harmful censorship and propaganda. Avoiding this might take careful attention. And since the scale of AI speech acts might grow dramatically over the coming decades, early truthfulness standards might be particularly important because of the precedents they set.