Bio-safety level 4 laboratories need to post cryptographic timestamps of all their records on blockchains such as the Bitcoin blockchain.
My qualifications: I have a Ph.D. in mathematics and have developed cryptocurrency technologies, but I personally do not have any medical or biological training.
Bio-safety level (BSL) 4 labs are capable of genetically modifying viruses that can infect and kill humans and other animals. BSL-4 labs therefore need to not only secure their facilities, but they also need to be investigable so that investigators can know exactly what is happening at all times in the BSL-4 labs even years after an incident takes place. The best way to do this is for the BSL-4 labs to not only collect extensive data on everything that is going on there but for the BSL-4 labs to also post cryptographic timestamps of all of this data on blockchains such as the Bitcoin blockchain. This procedure will discourage BSL-4 labs from doing dangerous research in the first place, disregarding safety precautions, and possibly intentionally causing harm.
To post a cryptographic timestamp, the BSL-4 labs can simply post cryptographic hashes of all of their data on a blockchain such as the Bitcoin blockchain. The BSL-4 lab may also post a Merkle root of this data on a blockchain so that it is possible for an entity to select and reveal specific portions of this data while concealing the remaining portions.
After a timestamp of a document has been posted on a cryptocurrency blockchain, the document cannot be changed without completely invalidating the timestamp. Given two points in time with , it will not be possible for an entity to produce a new document at time t with cryptographic timestamp pointing to time . A BSL-4 lab cannot change any information about any document that has been timestamped after the timestamp has been posted without invalidating the timestamp.
Here are a few advantages of posting cryptographic timestamps of records on blockchains:
Inexpensive: Computing cryptographic hashes or Merkle roots of data is very easy for modern computers. While it does cost cryptocurrency to post a transaction or information on blockchains, timestamping data on blockchains gets less and less expensive as more and more entities timestamp data. For example, suppose that it costs $10.00 to post a cryptographic timestamp on the Bitcoin blockchain. If 1000 entities need to post cryptographic timestamps of data on the Bitcoin blockchain, then it will still cost the $10.00 for all 1000 entities to post cryptographic timestamps of all their data together. This means that each of those entities will need to pay an average of $0.01 in transaction fees to post a cryptographic timestamp of their data on the Bitcoin blockchain. For all intents and purposes, posting cryptographic timestamps on the Bitcoin blockchain is free.
Easy to use: The BSL-4 lab only needs to post cryptographic timestamps of all of its data publicly and save the data since there will be (or at least there should be) web crawlers that search the web for cryptographic hashes and then post Merkle roots of those cryptographic hashes on blockchains. The web crawlers can also post a public database of all of these hashes so that it is easy to verify that these timestamps are correct.
Confidentiality and security: The cryptographic timestamps do not reveal any information about the data that is being timestamped.
Secure: Cryptographic timestamps on blockchains are more secure than the blockchains themselves. For example, suppose that a BSL-4 lab posts cryptographic timestamps of all of its records on an altcoin blockchain and the altcoin blockchain is being mined using 5 CPUs. Then there will most likely be (or at least there should be) a crawler that goes through all blockchains or web data, permanently records this data, and then posts the cryptographic timestamps of this data on the Bitcoin blockchain. Then these timestamps from the BSL-4 labs posted on the altcoin blockchain will be just as good as timestamps posted on the Bitcoin blockchain. Similarly, if a cryptographic timestamp is posted on a blockchain which does not use proof-of-work or even just posted on the internet, then that timestamp will be just as secure as if it were posted on the Bitcoin blockchain as long as there exists a crawler that posts hashes on the Bitcoin blockchain. The cryptographic timestamps are therefore more secure than any individual blockchain they are posted on.
Provable investigative report: When investigators investigate a lab incident, the report can prove that the information is legitimate. This will make it harder for even compromised investigators to produce a dishonest report on the lab incident.
BSL-4 labs may use artificial intelligence to genetically modify pathogens to increase infectivity and virulence. Since pathogens can reproduce at an exponential rate and evolve, these BSL-4 labs are capable of doing much harm. I therefore consider genetically modified pathogens to be a current AI related threat. Since evolution by natural selection is itself a sort of non-human intelligence that humans cannot simply turn off after it is released in the wild, we should consider dangerous pathogens to be similar to AI threats.
AI safety and blockchains: Cryptographic timestamps posted on blockchains can improve safety and investigability in many different contexts including AI safety. Suppose that an artificial agent were acting in the real world and not a simulation. Then such an agent can have as a part of its reward function a goal to post cryptographic timestamps of its data. If the artificial agent for some reason refuses to post cryptographic timestamps of all data, then other entities will know that the artificial agent is unsafe. I cannot think of a reason why an artificial agent that posts cryptographic timestamps of all of its data on blockchains will be less safe than an artificial agent that does not do this. And I can think of several reasons why an artificial agent that does publicly post these cryptographic timestamps will be safer than an artificial agent that does not.
Sadly, I am still the only entity who is publicly talking about how BSL-4 labs need to post cryptographic timestamps of all of their records (I searched the internet for posts about this issue made by others, but I was only able to find my own posts), and the only way I see this really changing is if the next pandemic possibly comes from a lab causes a few people to wake up and realize that BSL-4 labs currently have insufficient security standards. Since I do not expect for BSL-4 labs to post cryptographic timestamps of all their records on blockchains nor do I make anyone except for myself to make a significant effort in making sure that BSL-4 labs post these cryptographic timestamps, I have adapted my views about the security of BSL-4 labs.
I am now only asking for BSL-4 labs to publicly post cryptographic timestamps of all their records instead of worrying about making sure that these cryptographic timestamps make their way to public blockchains. For example, if a BSL-4 labs publicly posts a cryptographic hash or a Merkle root of all its new data every 10 minutes (on a website) without paying the transaction fee to put this cryptographic hash or Merkle root on a public blockchain, then that would be fine with me. In this case, BSL-4 labs do not need to pay cryptocurrency transaction fees, and people at BSL-4 labs can be completely oblivious to the usefulness of cryptocurrency technologies for bio-safety and that would be fine with me. But in this case, someone else needs to periodically take the cryptographic timestamps from the BSL-4 labs and post these on public blockchains. Perhaps a web spider can go on the websites for the BSL-4 labs and gather their cryptographic hashes to make sure that they end up on blockchains. Of course, in this case, for security, there should probably be multiple organizations that make sure that the cryptographic timestamps make their way to blockchains, but at least this will make it a little easier for BSL-4 labs to do what they need to do.
Someone on this site referred me to a paper with a similar proposal by those with medical and biological knowledge Aleksandr V. Kudriavtsev, Anna Vakhrusheva, and Alexander Shneider (but who have less expertise in information security and cryptography) which proposes using private blockchains (it looks like this is just a shared database where deleting information is not allowed) to track the origin and dissemination of scientific data. Their idea may have some merit (though such an idea would require more data to be shared between more people, and this could be a security hazard, so maybe it is best if just the timestamps are made public), but at this point, we need to implement the simplest and most inexpensive solution which is for the BSL-4 labs to post those cryptographic timestamps publicly.