The Opportunity Exists to Accelerate Science Today. Why Isn’t It Being Used?
Most people think scientists have access to the AI they need — the opposite is true.
When it comes to solving today’s most important and pressing problems, we should be using the most powerful tools available.
It’s well-known that artificial intelligence is a game-changing technology. Which groups of people have access to AI today?
We recently commissioned a small poll to ask people this question. Overwhelmingly, the response was “scientists,” “engineers,” and “researchers” working on problems like cancer.
It’s clear that a super majority of people believe that most scientists and researchers have access to AI.
For our respondents, as for most people, it was common sense the people working on problems like sustainable energy, clean water, or disease eradication would have access to these tools.
The reality is different:
In many industries today only 1 in 50 senior scientists have access to modern AI tools and software.¹
The small region of overlap in Fig. 1 suggests that by modernizing their toolkit scientists can be empowered to work faster, more efficiently, and to discover more.
The inability to access AI tools is not due to lack of interest from scientists. Instead, it’s because tools actually built for their use cases are few and far between.
The reality is many of the most widely-used scientific software programs today were originally written in Fortran code in the 1990s.²
In normal times the reason to upgrade tools for researchers would be to increase productivity with resulting lower overall costs. Today, the world is confronted urgently by COVID-19 and arming the majority of scientists with dramatically more powerful tools could help them turn the tide against the virus sooner.
So if it’s not the scientists, which groups of people are AI power-users today?
AI is vivid to us in our daily life. Search engines use data science techniques to deliver relevant search results. Shopping sites show you recommended products based on what you and other users bought in the past. Unsurprisingly, many data scientists work in consumer companies.
This is not to say people aren’t using AI for science. In fact, some of the most exciting discoveries of the past few years have come from applying AI — and we expect that trend will grow significantly.
There are indicators already suggesting an industry-wide shift towards using AI for science is underway.³ ⁴ ⁵ ⁶
That’s why making AI more accessible to scientists is a key part of our mission at Noble.⁷ ⁸ When it comes to COVID-19, many scientists who could be doing critical work are locked at home and need a more efficient way to do their research.
By building a community of scientists who are equipped to use AI for their own work, we have the opportunity to help drive science forward.
In today’s world there’s no time to waste.
¹ Noble survey of scientific organizations.
² Capterra: https://www.capterra.com/simulation-software/
⁶ Raghu, M., & Schmidt, E. (2020). A survey of deep learning for scientific discovery. arXiv:2003.11755. https://arxiv.org/abs/2003.11755.