Trust
How can we achieve trust with respect to digital information
How can we trust information that is gathered by people that are not ourselves without relying on their integrity
If someone wishes to fudge experimental results right now, it is fairly easy to do so. Simply open your CSV in an editor, change some numbers and now you have a dataset that is not accurate, yet presents identically to your peers.
We can extend this to more complicated information and more complicated editing tools. If you with to change or fabricate video and audio data, there exist many tools that allow you to do that and disguise that fact.
Much of the digital informaiton that exists today is not menually entered into a computer by a human, but created by a transducer from environmental readings.
We now have at our disposal modern cryptographic techniques that can guarantee certain information in certain contexts. How can we use these techniques to guarantee that information created by transducers can be trusted by even the people that were not involved in the experiment?
How to verify the legitimacy of information: transducers that turn environmental info to digital - in this case audio and visual information with cameras Certified footage and audio As it’s created, information is processed by custom hardware that generates an encrypted signature with a secure private key on hardware, transforming digital data that the transducer has captured This digital data can later be used with a public key to ensure that it has not been editied at all after being captured
After discussing with my friend Sam:
The system can never be perfect, so is it better to have more certainty in the entire data set, with some false content that is accepted as 100% true or is it better to have a distrobution of skepticism across the set of data based on factors that have been developed a long time (current system)? Even if the only result is raising the barrier of spoofing data, is this still valuable and productive?
The system falls apart anyway, even if implemented perfectly. Users know that no system is perfect and that there is a chance that a given piece of content that they are consuming is fake. Even if a crypto module is never broken into successfully and all verified content is actually true, if a user sees something unbelieveable, they will assume that the system has been hacked, rather than be convinced of the video
Just because the system falls apart, doesn’t mean that it’s not marketable. Is this ethical or does it just create more waste for no useful outcome?
Book idea: In Defense of Flat Earthers A book on information theory and how little we can be certain about events happening even a few degrees away from us Call to action in local communities Flat earthers, theoretically, are very good at questioning things and the essence of what they are doing is a good thing People say that you can very easily prove to yourself that the earth is round and not flat a la Galileo, but to be able to ‘trust’ this proof, you also need to trust everything it is built on, essentially requiring you to prove the entire history of mathematical and scientific proofs to yourself before taking a step forward in modern study Result: you must trust and make assumptions However, there are ways to mitigate error related to false trust See if this is a field of study within information theory and do a literature review
Introduce the concept of memory for humans (which everyone should be familiar with) - you can’t hold in your head all proofs leading up the the work that you are currently doing - there must be trust at some point.