Human keys secure scientific integrity

Disclosure: The views and opinions expressed here are solely those of the author and do not necessarily represent the views and opinions of crypto.news editorial.

The speed at which AI bypasses regulations poses risks to data, identity and reputation verification and, if left unchecked, could increase the prevalence of misinformation and slow the progress of scientific innovation. The march toward superintelligent artificial intelligence is represented by its most ardent leaders as progress toward a scientific golden age. But this push raises the possibility of existential risk that our society will hit a humiliating technology plateau, where the limits of immature AI technology become widely adopted and, over time, impair human creativity and innovation.

This is a contradictory approach for most accelerators. Artificial intelligence is expected to increase our ability to do work faster and synthesize greater amounts of information. But AI cannot replace inductive reasoning or the empirical process. Today, anyone can use AI to generate a scientific hypothesis and use it as input to create a scientific paper. The results of products like Aithor often appear reliable at first glance and can even pass peer review. This is a big problem because AI-generated texts are already presented as legitimate scientific findings and often contain fake fabricated data to support their claims. There is great incentive for young researchers to use whatever means they have to compete for the limited number of academic jobs and funding opportunities. The current incentive system in academia rewards those who can publish the most papers, whether or not the papers describe legitimate findings; they just need to pass peer review and receive sufficient citations.

Academic content with unverified authorship will also pose a significant problem for industries that depend on basic science to power their research and development; This is the R&D that keeps our society functioning and the quality of life of a growing global population. As a result, well-funded R&D can only rely on research it can perform and replicate on its own, increasing the value of trade secrets and dealing a devastating blow to open science and access to meaningful information.

Expensive replication efforts may overcome misinformation on their own, but the problem is much bigger than that. Today, we face an erosion of trust in the foundations of knowledge, where unverifiable claims and vague attributions undermine scientific advances and pose a threat to the scientific community. There is an urgent need to establish a fact-based economy for reliable verification of content and data.

AI systems are as powerful as the data they are trained on

Large language models are excellent tools for creating persuasive content; but they are only as informative as the data on which they are trained. Their ability to make predictions outside of the training set is still limited. The role of science is not only to synthesize existing knowledge, but also to create new informative works that increase the entropy of the collective corpus of knowledge accumulated by humanity. Over time, as more people use AI to produce content and fewer people produce original content, we will face a “low-entropy bloat” that does not present new information to the world, but merely recombines past information. Unless we build a flexible source and verified attribution layer into the AI ​​tools used for serious research, primary sources will be lost as new “knowledge” relies on self-referential AI-generated content.

This “lobotomization” of the intellectual depth of the collective human community will lead to lasting effects on medical, economic, and academic research, as well as on the arts and creative pursuits. Unverified data can influence studies, skew results, and lead to significant policy or technology failures that erode the authority of scientific research. The risks of AI-generated “science” are manifold. The ordinary course of normal science will come to a halt due to authorship disputes, allegations of plagiarism, and corruption of peer review. We will need to devote more time and energy to dealing with the many consequences of declining quality and accuracy of scientific research.

Artificial intelligence is a useful tool for sparking ideas, structuring thoughts, and automating repetitive tasks; It should remain a complement to man-made content, not replace it. It should not be used to write scientific papers proposing original findings without doing the work, but rather to help improve the efficiency and accuracy of human-led efforts. For example, AI can be helpful in running simulations on existing data using already known methods and automating this work to help explore new research directions. However, the experimental protocol and human creativity necessary for scientific research cannot be easily replaced.

Building a truth-based economy

A fact-based economy creates a framework of systems and standards to ensure accuracy, integrity, transparency and traceability of information and data. It addresses the need to build trust and verifiability across the technological community by allowing individuals and organizations to trust the accuracy of shared information. Value is based on the accuracy of claims and the authenticity of observations and primary sources. A fact-based economy will make digital information “hard” just like Bitcoin made fiat hard. This is the promise of the decentralized science movement.

How do we get there? We need to start with the individual researcher and their work, which is the most important element of the scientific world. Today, existing web standards for scientific identity fall short of verifying claims regarding identity and proof of work. Current practice makes it very easy to produce a profile with an acceptable reputation; Peer reviews are also at risk from bias and collusion. A fact-based economy for science cannot be established without verifying the metadata that accompanies a scientific claim.

Improving academic identity standards can start with a simple cross-platform login supported by privacy-preserving authentication technology. Users must be able to log in to any site with credentials, prove authenticity, and selectively disclose reputation, data, or facts about other agents or users.

A layer of identity based on a verifiable researcher’s reputation forms the fundamental foundation of DeSci. A complete on-chain scientific economy would allow public and anonymous participation in massive online coordination of research activities. Research laboratories and decentralized autonomous organizations can create permissionless systems and reward programs that cannot be fooled by false claims of reputation or identity. A universal scientific record secured on the blockchain with identity claims would provide a frame of reference for autonomous organizations created to collect verifiable scientific information and test falsifiable hypotheses.

Securing the future of human progress

To prevent the breakdown of trust in specialist areas of research, we need to establish foundations of truth through information transparency and rigorous verification. The chances of our collective progress continuing for hundreds more years and paving the way for successive scientific revolutions in materials science, biotechnology, neuroscience, and complexity science will depend on quality research and the curation of robust data. This will be the difference between a future society as advanced as ours compared to pre-Enlightenment societies. Otherwise, we as a species will have to wait for how smart this is and get even stupider. It’s unclear if DeSci will save us, but there’s a limited amount of time to get things back on track.

Shadowy El Damaty

Shady El Damaty is the co-founder of the Holonym Foundation, which seeks a solution for universal personhood and secure digital access with a decentralized identity protocol built on the magic of zero-knowledge proofs. In 2020, he founded OpSci, the first decentralized science, or DeSci for short, organization. Prior to his career in crypto, Shady received his PhD in neuroscience from Georgetown University in Washington, DC, USA.