About the CRS

In late 2018, a small group of individuals with backgrounds in computer science, psychology, and linguistics started analyzing the state of artificial intelligence. As our launch manifesto explained, we disagreed heavily with the approach that current companies and investigators were taking with respect to the development of next-generation artificial intelligence.

Our position may be summarized as three individual points:

1. The most objective, measurable evidence of advanced capability by living organisms seems to be that of human beings. For example, within a period of 150 years, we went from horse and buggy to the moon, genes to machines, and now we are trying to go from the Internet to artificial intelligence. No other living organism that we know of has accomplished such feats in such a short amount of time.
2. The mechanism by which advanced human capability works is poorly understood. For example, there are competing schools of thought in the field of psychology, and there is no grand unified theory that can explain all of human history. This is why prominent scholars disagree on basic definitions for words like consciousness, intelligence, and learning.
3. Without clear definitions for the words and concepts outlined above, it is improbable, if not impossible, to replicate human ingenuity to higher orders of magnitude. For example, if one is building artificial intelligence, but simultaneously cannot provide a precise definition of "intelligence," then there is no clear, measurable goal.

The implications of these three points are legion. It means that no amount of money, resources, or time will ever result in artificial intelligence orders of magnitude greater than human capability. In the computer science community, there is remarkably little discussion about this pink elephant in the room. The statements above may be regarded as our contrarian opinion.

We believed then, as we believe now, that in order for humanity to achieve a different result, we have to change our behavior. This requires reassessing core assumptions about human capability, computer science, and a number of other poorly understood topics. In early 2019, we launched the Consciousness Research Society (CRS) under the domain name conscious.ai in order to investigate the assertions we outlined.

Since 2019, The CRS adopted a three-step plan with a long-term view. We believe that a tangible path to next-generation artificial intelligence requires defining "intelligence," developing novel theory, and then building the technology. While most computer scientists are thinking about the future, we are uniquely doing the opposite. That is, our starting point requires us to go back in time to the earliest human writings in order to evaluate human mentality. This means studying ancient texts, languages, and cultures.

From 2019–2023, we have invested heavily in research projects focused on understanding human mentality. All of our research is shared with the public for free.

As of July 2023, we are completing the first of our three phases: defining important terms like "intelligence" with precision. Interestingly enough, artificial intelligence is gaining mainstream attention as a result of technologies like ChatGPT by OpenAI, Bing by Microsoft, Bard by Google, and countless others. As our forthcoming open letter will explain in more detail, those technologies are not examples of artificial intelligence, but artificial learning. Common viewpoints in the field of computer science do not seem to distinguish between the two.

As noted in our 2019 launch manifesto, there is still mass confusion about the precise definitions for these important terms. This is not to discredit the incredible technological advancements made available to the public. Nevertheless, we hold that artificial learning is a feature of artificial intelligence, not the product itself. Indexing the Internet and providing a conversational interface for questions and answers does not represent intellectual capability orders of magnitude in comparison to human capability. Let us not forget the high standard we have set. The best chatbot can summarize a cooking recipe or write snippets of code, but it cannot find a way to go to the moon and back. Only human beings have accomplished the latter. The importance and objectivity of this example cannot be stressed enough.

For the remainder of 2023, the CRS is working on a comprehensive framework for analyzing human behavior in the form-factor of writing. The working name of this project is called PsychLing. This foundational step will include preliminary patent applications for several critical technologies that we will be testing in the second phase of our research enterprise.