# Introductory notes on biorisk <div class="pills-container"> <span class="pill">Published: January 29, 2024</span> <span class="pill">Reading Time: 5 minutes</span> <span class="pill">This is a linkpost to <a href="https://halfbakedtheories.substack.com/p/20240129-3">Substack</a></span>. </div> The [Global Challenges Project](https://www.globalchallengesproject.org/) was the first time I immersed myself into reading up on bio-risks. This article on the [on-ramps model into biosecurity](https://forum.effectivealtruism.org/posts/QeEgktwh2FQyot9Jw/on-ramps-into-biosecurity-a-model) was shared to me and might be useful for people who are considering a career on this field. See this [map of biosecurity interventions](https://forum.effectivealtruism.org/posts/npNt43QRnaRNRixXK/map-of-biosecurity-interventions) for a view of what people are working on so far. ## What are global catastrophic bio-risks (GCBRs)? Global catastrophic bio-risks (GCBRs) are broad biological threats, like pandemics and bioweapons. They have the potential to cause worldwide damage and high mortality. Specifically, they have the potential to kill about 10% of the global population and cause human extinction or an irreversible collapse of civilization to be considered an x-risk. Historically, this has happened before. [The Black Death](https://en.wikipedia.org/wiki/Black_Death) showed how much of a devastating impact such a risk could do to our society as it killed off around 10% of the global population. With advancements in technology, there is a growing concern about artificially created pandemics which can be designed to both highly contagious and virulent. As mentioned earlier in the article, AI can be a dual-risk since modern technology increases the possibility of creating more pathogens. There is also a risk of accidental or intentional release of engineered pathogens. In the past, handwashing, antibiotics, vaccines, and even international treaties have been good enough countermeasures to reducing GCBRs. Today, research on early warning systems, PPE, better facilities, and ensuring smart governance has been the priority within the field. Some other countermeasures for bioweapons are: 1. **Switching:** Ability to quickly change or adapt the type of countermeasures being used in response to different bioweapons. 2. **Escalating:** Increasing the level of response as threat level escalates. 3. **Attending hazards:** Identifying the most pressing threats and concentrating efforts on mitigating those specific dangers. Unlike AI, the risk in bio-risk isn’t solvable with more time. Within a pandemic or a bio-risk outbreak, you are fighting with time. Your best hope would be to ensure that the systems and interventions build to prevent the outbreak from happening in the first place or at least to mitigate its impact would actually work. Unlike AI, bio-risk directly kills people and societies. As part of the countermeasures being prioritized right now are managing substitution effects like understanding the consequences of certain actions. This means trying to avoid triggering an arms race in defense technologies and (ironically) ensuring confidence in winning such a race. You wouldn’t want a bioweapons war, but if there was such a war, ideally you want to win and stay alive. ## How do you get involved in mitigating bio-risk? Interestingly, most work on mitigating long-term bio-risks can also help in the short-term. An approach could be developing mathematical models to predict the spread of diseases or explain disease dynamics, forecast future trends, and provide a sandbox for policy levers. However, a lot of mathematical models don’t have biological reality in mind. To address this, in the past, social media has been used to incorporate the social aspect of this reality. (Meta-)genomic pathogen surveillance is also a very helpful way to engage in bio-risk. [Metagenomics](https://en.wikipedia.org/wiki/Metagenomics) is the study of genetic material that is recovered directly from environmental samples. To use metagenomic techniques in pathogen surveillance means to monitor and analyze pathogens by looking at samples extracted from nature. Other ways to get involved would be: - Establishing centers for early detection of biological threats. - Researching and developing new sterilization technologies based on physical principles. - Developing superior personal protective equipment. - Enhancing the Biological Weapons Convention to better prevent the development and spread of bioweapons. ## AI risks and bio-risks are intertwined The Oxford workshop was the first time I heard of the problem of the AI and bio dual-risk. This is actually fairly straightforward. Better technologies means more sophisticated future biological threats. Better LLMs can also provide information on how to misuse biology without needing an extensive biology background. Read “[Can large language models democratize access to dual-use biotechnology?](https://arxiv.org/abs/2306.03809)” and “[Will releasing the wights of future large language models grant widespread access to pandemic agents?](https://arxiv.org/abs/2310.18233)” to look more into dual-risks. ## On building a career towards reducing existential risks (x-risks) College can be inefficient and slow. I’ve learned over time that the best way to learn is to leave open the possibility of intensely skilling up or studying a new focus area. This means reserving time and energy towards activities that matters more to where you want to be in order to maximize your impact. For your own views on what needs to be done. Having mere conversations is not a step towards this. Researching and articulating your thoughts aside from these conversations are more valuable because it takes a rich understanding of the problem to be successful. Build your knowledge my reading and build your attitude by taking action. A speaker emphasized the importance of certain underrated priorities, especially when engaging in the academe. To them, their biggest career mistake so far was not being strategic with doing their PhD. They weren’t clean about what they wanted to get out of it. They didn’t work out what to prioritize based on their goals. They didn’t take the downsides of PhD very seriously. Lastly, they didn’t notice that things were already going badly. As a result, they were distracted by pursuing local and minor incentives. They got stuck in a loop where they feel bad about work, then they procrastinate, which makes their work quality much worse, then again feel bad about it. Overall, they lost a lot of time. As advice, they said that being strategic and considering avoiding failure modes should be part of a researcher’s priority. That is: - Prioritize understanding yourself (i.e., motivations, relative strengths, weaknesses, countermeasures) - Prioritize looking after yourself - Prioritize learning about the things that you need to understand (and identifying these blockers as early as possible) - Prioritize having a scout mindset (look for ways you are wrong and not ways you are right) - Make sure your self-worth isn’t just in your impact - Spend time with people you want to be more like (you are the average of the people you spend the most time with) - Set up a reflection process that works for you