[ad_1]
Thanks to widespread access to food and medicine, travel, and many other scientific and technological developments, billions of people today are better off than the kings of the past. I am very surprised that this amazing fact is not appreciated more.
Of course, despite all the progress we have made, the world faces many challenges in the 21st century. climate change, pandemics, poverty, and cancer, just to name a few.
These challenges could be addressed and overcome if all countries of the world joined forces and shared technology and resources. However, there are no examples of such cooperation in history, and the current geopolitical situation does not offer much hope.
Our best hope for meeting these challenges is to increase scientific and technological productivity. The only viable way to achieve this is by integrating artificial intelligence (AI) and laboratory automation.
AI systems already have superhuman scientific capabilities. They can memorize vast amounts of facts and learn from huge datasets. They can perform perfect logical reasoning and near-optimal probabilistic reasoning. They can read every scientific paper, in fact every paper ever written. These powers are complementary to human scientists.
When the scientific method was developed in the 17th century, one of the core insights was the need to conduct experiments in the physical world, not just think.
Currently, laboratory automation is steadily advancing, and robots can now perform most of the laboratory tasks that humans can perform. Additionally, the concept of a “cloud lab” is now emerging. The idea is to provide laboratory automation at scale and remotely, where scientists send samples to a cloud lab and design and run experiments using a computer interface.
And then there are the AI scientists. It is an AI system integrated with laboratory automation that can perform closed-loop automation of scientific research (also known as “robot scientist”, “self-driving laboratory”). These systems automatically generate hypotheses to explain observations, devise experiments to test these hypotheses, use laboratory robots to physically perform these experiments, and evaluate the results. Interpret and repeat the cycle.
AI scientists can work cheaper, faster, more accurately, and longer than humans. You can easily increase it. Because experiments are automatically devised and executed by computers, every aspect of the scientific process can be fully captured and managed digitally, increasing the reproducibility of science. There are currently around 100 of his AI scientists around the world, working in fields ranging from quantum mechanics to astronomy, chemistry to medicine.
Over the past year or so, the world has been stunned by the success of large-scale language models (LLMs) such as ChatGPT. LLM achieved breakthrough performance on a wide range of conversation-based tasks. LLM is incredibly powerful in absorbing technical knowledge such as chemical reactions and logical formulas. The LLM, and more broadly the Foundation model, shows great potential to strongly support AI scientists. They act as a source of scientific knowledge as well as a source of new scientific hypotheses as they read all the scientific literature.
One of the current problems with LLMs is their tendency to hallucinate, or output statements that are not true. This is a serious problem in many applications, but not necessarily in science, where physical experiments determine truth. A hallucination is a hypothesis.
AI has been used as a research tool behind tens of thousands of scientific papers. We believe this is just the beginning. We believe that AI has the potential to transform the scientific process itself.
We believe that by harnessing the power of AI, we can propel humanity into a future where breakthroughs in science, even Nobel Prize-worthy achievements, can be fully automated. Such advances could transform science and technology and offer hope for meeting the formidable challenges facing humanity in the 21st century.
By 2050, the Nobel Turing Challenge aims to develop AI scientists who can make Nobel-quality scientific discoveries on par with, or better than, the best human scientists.
While the application of AI to science is a potentially transformative force for good, it also has the potential to cause harm. As a step toward preventing this harm, my colleagues and I created the Stockholm Declaration on AI for Science. This commits signatories to the responsible and ethical development of scientific AI. You can sign a copy of the declaration at https://sites.google.com/view/stockholm-declaration.
We urge all scientists researching AI to sign.
[ad_2]
Source link
