Empire AI is Essential Lifeline for Academic Research
When it comes to solving the internet’s most dangerous problems, time is truly of the essence. Things like online extremism and harassment campaigns can quickly take on new forms, and it’s important to have the technology necessary to adapt on the fly.
Fortunately, New York has given us such a tool: Empire AI, which will soon be the most powerful academic research computer in the country. Especially in light of the federal government’s decision to pull back on supporting academic research and science, New York’s Empire AI has gone from a competitive advantage to an essential lifeline for researchers like me.
I have dedicated my career to studying jerks online; how much harm they do, and how we can overcome their cruelty. The rise of the internet fundamentally changed our world, giving us advantages we wouldn’t have thought possible twenty years ago. Instant communication has eliminated geographical barriers that have historically restricted access to information, culture, and science. But with these advances have come with serious costs, including a rise in disinformation, harassment campaigns, extremism, risks to children, and antisemitism, which is a focus of my research.
Artificial Intelligence (AI) is an integral part of my work, but it’s also given criminals and other wrongdoers more power to sow discord and hate. I use AI to learn about online human behavior, but miscreants use AI as a force multiplier for their dangerous purposes.
Empire AI gives me the upper hand as I combat these dangerous forces.
One year ago, it was just an idea, proposed by Governor Kathy Hochul and approved by the state legislature in April. With significant support from the state, Empire AI is a consortium of public and private universities in New York: CUNY and SUNY – where I work – but also Cornell, Columbia, NYU and RPI.
Originally, our goal was to build a state-of-the-art artificial intelligence computing center at University at Buffalo, scheduled to open in 2026. But in October, Empire AI launched Phase Alpha, a smaller, but still incredibly powerful version of what will be the final supercomputer. Alpha lets researchers like me access the 245th most powerful computer in the world, including 96 of the most powerful computer processors currently available for training AI models, 18 months ahead of schedule.
The work currently being conducted in my lab involves understanding how AI can be used to realize policies related to issues like hate speech. To do this, I leverage AI to gain a large-scale, quantitative understanding of these problems. Using our old hardware, it was impractical to analyze the billions of data points my work explores.
My research can now be done at a speed I would have never had access to before. Previously, it would take nearly 20 years to run our experiments on even a sample of the billions of social media data points we collect yearly. But by pooling resources from New York State and the six member universities to build something more powerful, I can complete my research of online antisemitism in a matter of weeks. It is hard to overstate just how much more this increase in computing power will allow me to accomplish.
The same hurdles Empire AI will remove from my work hold true for others. The Alpha computer is already running at maximum capacity, with 85 projects and more than 250 researchers across the state using it to do work for the public good, including protecting power systems from malicious attacks and developing a robotic seeing-eye dog for people who are visually impaired, research being done by my colleagues at Binghamton University.
Empire AI will also open unique doors for collaboration between disciplines. Scientists and faculty in the humanities and the arts will have expanded access to computing power that was previously unavailable to public research institutions. This coalition will open doors to scholars, especially to those in public universities dedicated to establishing New York state as a leader in AI.
This is a pioneering effort to ensure researchers have access to the computational power necessary to understand and ethically deploy AI for the public good. Empire AI allows us to harness the positive power of the internet, while also fighting against some of its worst dangers. Now is not the time to back away from critical academic research. Thanks to New York State, we don’t have to.
Jeremy Blackburn joined the School of Computing at Binghamton University in fall 2019. Jeremy is broadly interested in data science, focusing on large-scale measurements and modeling. His largest line of work is in understanding jerks on the Internet. His research into understanding toxic behavior, hate speech, and fringe and extremist Web communities has been covered in the press by The Washington Post, the New York Times, The Atlantic, The Wall Street Journal, the BBC and New Scientist, among others.
