 My name is Tom Dixon, and I'm the vice president for the Australian Institute of International Affairs in New South Wales. I'm also the manager of National Security and Defence at Macquarie University in Sydney, and it's at Macquarie that I'm undertaking a multi-disciplinary PhD investigating the impact of engineering biology on international relations security and defence. In this presentation, I'm going to discuss the coming age of robust two-way communication between living and non-living systems. This is something that a co-author and colleague of mine at Macquarie, Dr Tom Williams, has called the Internet of Biological Things. This is Tomorrow's World, a world arising from advances in optogenetic, bioelectrical and biosense-mediated communication systems. These are the three increasingly well-defined substrates through information management and communication in living systems, like electricity and chemicals, or perhaps in a more fundamental level, photons, electrons and a material structure of matter. Let's call them the building blocks of information's physical instantiation in reality, and this is what I want to make my remarks about. That life is information instantiated in reality, and therefore living systems inherit all the strengths and weaknesses we know so well about today's digital optoelectronic systems. Now in part, you're watching this because you already know about these weaknesses, and you know many of the emerging issues well, far better than I. Where I want to go in this talk is to openly discuss the need for a platform-side anticipating emerging security vulnerabilities arising from this transformative moment, and this is what I mean when I use the word cyber-biosecurity. I'm not talking about the protection of biological information when stored in inanimate substrates, as important as that issue is. Rather, I'm focused on the issues that are going to emerge from information systems where the information can transit between a living organism and an inanimate device, where computers can actuate biological functionality with low latency, and where biological organisms can actuate inanimate statistical systems with similar latency. Now I've often heard said about research interfacing neurological substrates with optoelectronic devices, that a door once opened can be walked through both ways. My talk today is about that door, and thinking about the impact that door might have on our world. That thinking has to happen now, before the door has been opened. Let me begin by sharing my screen. This is the puzzle, a simplified map of reality that outlines the issue I want to address in this talk. This map covers over length scales on our planet from small to large. On the left it captures a small selection of animate information substrates, and on the right a small selection of inanimate substrates. At the bottom is a simplification of the building blocks. In hindsight, I would have included RNA, DNA and humans on the left, and I wouldn't have narrowed the electromagnetic spectrum to visible light waves, but I hope the image makes its point. There is just as much, if not more, information naturally occur on the left hand side of this map than there is on the right side, and that's the puzzle. Now if I separate out the concepts behind the Internet of Things, the Internet of Everything and the Internet of Biological Things, this map becomes quite useful. The potential of two-way communication between living and non-living systems mediated by optogenetic, bio-electrical and biosensor based systems means that we need to reevaluate the future of networked information systems. Till now, information has primarily traveled one way on this map. Information has been extracted from living systems into inanimate systems, where it has been analyzed, manipulated and stored. The security of those endpoint inanimate systems has been a prime point of concern. No one seriously countens the possibility of hacking those inanimate systems by compromising the living systems prior to information harvest. However, it is something entirely more reasonable to suggest that information that's about to be communicated from the right hand side of this map over to the left hand side could be compromised prior to transfer. Those moments when information is still dished for the instantiated comprised of potential vulnerability. The mRNA vaccines are a perfect case study for this. Designed in silico, manufactured via cyber systems, injected into humans to hatch vulnerability within the mammalian organism. Though the time scale might be long compared to digital information system standards, the pattern is clear. The era of two way information communication between living and non-living systems has arrived and 2020 marks the point of no return. I think about the mRNA vaccination process is similar to analog computing engineering prior to the 1950s. It is just the beginning. In the 1940s and 50s, how many people accurately predicted any of the 21st century's emergent security issues that have arisen from the progress of optoelectronic engineering and the information and computing sciences. We are completely out of our depth when it comes to predicting anything about types of security issues on the horizon. And these security issues are going to be specific to bioinformational engineering. That is engineering integration of cyber and biological information systems. Hence there is an urgent need for horizon scanning of a new kind on a new scale. The key question being how do we anticipate new and normal threats arising from the internet of biological things, especially when we have little to no idea what an IVT enabled world will actually look like. The canary in the coal mine for this transition to a bio informationally enabled world is the convergence of synthetic biology and artificial intelligence. Now I'm fascinated by this convergence because of its value in representing the current reality, but also because of its symbolic, moral, ethical and philosophical value. Now if we move past the identification of risks, threats and vulnerabilities for a moment and consider that new to the world organisms are being designed by algorithms. What does that even mean? I've closely followed Delphi scans run by the Centre for the Study of Existential Risk at Cambridge University. They've run two scans on issues arising from advances in engineering biology and I recommend both of them to you. Their scanning technique is based on a model used to identify issues in conservation and ecology with a decade of use behind it. And that Delphi process in turn is based on a model that was pioneered by US Defence in the mid 20th century. That one was created to anticipate technological surprise. And that's where we are again today, living with an urgent need to anticipate technological surprise. Without digressing too far, another work I recommend to you is Future Shop by Alvin Toffler published in 1970 is an extremely pressing work. For many the future has arrived too soon and this experience of the future arriving unexpectedly is particularly pronounced among policy makers since COVID-19 reshaped the world. COVID-19 was a biological event predicted by good many experts and pandemics that sat on international and domestic watch lists for decades. Yet it's still surprised many when COVID-19 occurred and many are still suffering Future Shop because of it. My contention is a simple one. Natural origin biological events, though they are the most likely biological event to cause catastrophes are not the only type of biological event out there. Engineered biological weapons entered the playbook sometime during the 20th century and there they have suddenly remained as high risk high impact low probability events. But bioinformational engineering has the potential to create dozens if not hundreds of new types of biological events for which we don't even have words to describe them right now. I would label this black box as an area of unknown stochastic biological shocks. Connecting information systems of the living world to the non living world is going to vastly increase the potential for stochastic biological shocks. Why? Because we're creating closed and open control loops between information systems in the biological world in the digital. As we know all too well from the digital world. Critical issues in information systems can scale logarithmically. Are we prepared for new kinds of biological shocks to scale logarithmically throughout the global biome? Of course not. I would suggest that we need to urgently begin preparing for new types of stochastic biological shocks, be they natural, accidental or engineered in origin. This means more horizon scanning, constantly iterating innovations in applied foresight and integrating those findings into policy and practice in a preventative manner on a much faster time scale than occurs today. Now this sketch is of a one way information process involving a Sentinel plant. And I'm rather fascinated by Sentinel plants. For those who haven't followed their development closely, their natural or transgenic organisms planted among crop field and closely monitored because they provide early warning of various natural processes being experienced by the surrounding crops. Be it heat stress, water stress, water stress, lack of nitrogen or phosphate, they can also monitor for pest and blight outbreaks. Now at the moment we're going through a transition. Sentinel plants have typically been monitored by humans or drones at length scales visible to human eye. With the invention and development of some next generation chemical detection systems, think diabetic monitors are plants. The biochemical signaling mechanisms of a Sentinel plant can be hardwired into psychophysical system control loops. This is still a one way loop. But I want to emphasize that over the next decade, Sentinel plants are very likely to be incorporating micro and nano scale information export systems that means signals that aren't visible on the centimeter scale can be monitored. One of these signals I'm particularly interested in is the immune system reaction of plants to unknown pathogens that might have been transmitted to them by nearby mammals grazing on foliage. Indeed, one of the more out there proposals I've made with some of my co-authors is that we should consider using Sentinel plants in the wastewater streams of avatars in order to monitor for disease X encounters. Let's get in front of the next zoonotic biological event rather than wait for it to occur. But the second point I want to make about this control loop is that there are many different points where the biological information is transiting an inanimate substrate. This point is totally vulnerable to standard cyber issues. However, if the inner workings of the Sentinel plant are being monitored by a device in a crop field and that device itself is not being monitored regularly by drones or humans, then compromising the hardware at the Sentinel plant communication system can offer a pathway into all of the linked systems. I'd rate that as an unlikely event for a crop field. But if we translate the scenario into a military context and Sentinel plant is monitoring some molecular signatures associated with an adversary's troops assets or chemical biological and radiological weapons, then the risk and consequences of hardware being spoofed, hacked or compromised rises. The base even persistent living monitoring systems that target micro and nano scale information signatures are disruptive technology. They're likely to be destructive from a military standpoint in the maritime and land domains, but they're also likely to be destructive from a civilian standpoint because their proliferation throughout urban rule and wild environments seem somewhat likely over the coming decades. Many environmental monitoring regulations can be automated by these systems, which may make their uptake by mining and resources extraction companies likely. Similarly, many environmental monitoring processes in homes and cities can be automated this way, meaning air quality in the presence of viral particles in the air can be monitored in close to real time. And even if we remove the plant entirely, the training towards the molecular monitoring of the planet is on and it needs to be prepared for. But again, this is only a one way control loop. We're not yet actuating biological functionality from an inanimate substrate. Now this is a closed loop by two way bioinformational system. My opinion is that innovations in this kind of system are going to contribute to scaling up bioreactive based by a manufacturing. As probably an inherent limit and how big a bioreactor can get before the underlying biomass will start to deviate from its intended functionality, but optogenetic and bio electrical controls may be able to contribute to increasing the edge of that limit. Perhaps more importantly, these types of information based controls are very likely to optimize production T to rates and yields. This is a gateway technology for gaseous carbon reuse and the inexorable movement of global manufacturing to a carbon negative bio economy. This image underpins the potential of bio manufacturing in the 21st century for me, especially when it comes to circular economy tools for mitigating and adapting to climate change. However, this system is also a textbook example of a cyber biological interface. And though it is a closed loop interface, it is clearly a hybrid surface that opportunists and advanced persistent threats will seek to navigate. Now I'm not going to recite to you the anticipated economic bounty of tomorrow's bio economy, but the forecast all point one way. The companies that are going to use these systems or relying on these systems in their supply chain or in their waste chain are all going to share something in common, the cyber biosecurity risks that arise from reliance on cyber biological interfaces. Now I've gone quite deep quite fast on this example and so I want to relook back to a singular question. How is a non technical policy and understand and interact with the security risks of the system, let alone the economic potential. The answer, we need to find a new language for describing the world encompassed by this emerging technology, a language that is sufficiently complex to adequately describe the underlying scientific and technological reality, and a language that is simple enough so that policy makers can interpret and understand the opportunities and challenges at hand. Where is that language going to come from. Well, in many ways, I think we can learn a lot from the path of artificial intelligence over the past decade. I recall when the AI hype cycle was really setting off a few years ago. And every second week there was an opinion piece and the content marking outlets of the McKinsey's and Bane's of the world, describing how AI was important and why business leaders need to pay attention. Each one of these articles though it didn't contribute that much refined and guided popular culture understanding of AI's possibilities and the language used to describe those possibilities. We need this to happen all over again to the dawning age of the cyber biological interface, because the cyber bio interfaces fundamentally different to a cyber physical system. As you know, a cyber physical system doesn't monitor manipulate or interact with biological information substrates in the material reality it is manufacturing. Cyber physical systems do not transfer information between animate and inanimate matter. Cyber bio interfaces do. And though these interfaces are not all that common today, they do exist and they are proliferating. Cyber bio novelty is going to take many by surprise, but it doesn't need to. Now, before I conclude this presentation, I want to make a few remarks on anticipating cyber biotechnology surprise and the importance of cyber biosecurity. This is my final slide. And I'm going to conclude with it. This is far more than a hypothetical situation, but it is one that we need to actively prepare for. This sketch is based on a 2016 proof of concept study where humans wearing EEG devices linked up to an optogenetic system in a mouse model could activate engineered E. coli within that mouse model to express insulin just by thinking about it. This is the world of bioinformational engineering. We're headed towards where medication can be switched on by devices by optogenetic control loops actuating functionality and engineered microbiomes and this is just one example. I'm sure you can let your imagination run a little bit wild on how important it is going to be to secure this cyber bioservices from opportunists and APTs. This is the cyber biosecurity conversation we need to be having. This is the dual use research conversation we need to be having. Perhaps most importantly, the medtech startups of the world that are going to pioneer this space all need to be brought into a cyber biosecurity first wave thinking without anchoring their limited venture capital runway. Now I fully acknowledged some of the sketches I've shown in this presentation may never eventually reality, but the science and technology is sufficiently advanced that bioinformational engineering is going to begin appearing in commercial applications over the coming decade. And there is an obvious first mover advantage for those companies and nations that pioneer this technology, especially from a standard setting point of view. That includes data standardization, but also the standardization of security protocols or creating long term structural weaknesses in the industry standard security settings for a bioinformational communication process. Now from my perspective, this is what cyber biosecurity is all about anticipating and repairing for the transformative potential of the cyber biological interface. And figuring out how to navigate the attack service that that is going to create. It's been an absolute pleasure to present to you. Thank you very much for listening.