Security Review: Robot Scientist automates the discovery of new drugs!
A group of researchers working in the EU funded IQ project have put together a robot scientist that can devise a theory, come up with experiments to test the theory, carry out the physical experiments, interpret the results and then repeat the cycle. This computer system has the potential to revolutionize drug discovery which until now was either serendipitous or was the result of random experiments on thousands of chemical compounds. This technology which employs advanced artificial intelligence and data mining could pave the way for more efficient, cheap, effective and speedy methods in areas of scientific research where there is some level of automation required.
Development of medication is a sensitive topic as it is vital to medicine, the health of all people and also because it has core political and economic implications. So in order to prevent customer abuse, many governments regulate the manufacturing, sale and development of drugs. In the case of the robot scientist, a rough approach towards finding a drug has been made easy and hence it is important to analyze its security. The major components of the system involve an inductive database which stores all the raw data (chemical compound compositions and pharmacological activity), data patterns and results of experiments and the artificial intelligence software and the experiment conducting physical portion.
Assets/Security Goal:
§ A major asset and security goal of the Scientific Robot which uses data mining is to protect the integrity of its database which houses all the raw data, patterns and experiment results and prevents them from falling into the hands of adversaries.
§ Another important security goal is to make sure that the robot that handles all the fluids and compounds and performs the experiments, is in good working order. Even if everything else in the system is fine, if any of the robot’s parts are sabotaged, then the experiment results would be incorrect, misleading and of no use.
Potential Threats/Adversaries
§ A possible adversary could be a competing pharmaceuticals company that wants to take advantage of the data and results collected by the robot up till now and use the stolen information for their own research. The motive could also be just jeopardizing the company’s prospects in medicinal research by modifying the database and hence affecting future experiment results.
§ Another threat could be from a malicious third party that wants to create medical havoc and achieves that by extracting information from the database about the compounds that have an adverse effect on the human body to concoct harmful drugs.
Potential Weaknesses
§ A weakness could be introduced into the system by implementation bugs in the data mining software and the AI software. This could lead the robot to make wrong decisions as to which experiment to conduct and all the following results which infer this will be wrong as well.
§ Another weakness is having all the data stored in one database. In the light of some technical issues or power outages, there is a risk in losing all of this carefully gathered and computed data.
§ If access to the database is not authorized, this makes the system vulnerable to an attacker who gains access and manipulates its contents.
§ The physical component (the liquid mixing robot) which actually performs the experiments is also the cause for another weakness. A part loosened or worn out, a screw missing here and there could mess up the outcomes of an experiment.
Defenses
§ It is necessary to have restrictive authorized access to the database so that information does not fall into the wrong hands.
§ There should also be a backup for the database to prevent data loss in case of emergencies.
§ The physical components should be checked for maintenance issues from time to time and be monitored to some extent.
Overall, this scientific robot is going to be a great asset to medicinal research by improvising greatly the current procedures used. It is important for us to understand that this system is dealing with and storing highly complex, sophisticated data which has a close connection to the future of medicine. Everything feasible should be done to make sure that this system is used for the benefit of human kind and not otherwise.
http://www.sciencedaily.com/releases/2009/02/090202140042.htm