ALBANY COUNTY — “This could be our Sputnik moment,” Assemblywoman Patricia Fahy told an audience that gathered last Friday at the University at Albany to hear a panel of speakers discuss artificial intelligence and automation.

She was referencing the Soviet Union’s 1957 launch of the Earth’s first artificial satellite. This jolted the United States into the space race. Currently, Fahy said, the U.S. is far behind China when it comes to AI investment.

Few terms these days hold connotations as paradoxical as “artificial intelligence” and “automation” do — except maybe “electricity” and “industrial revolution” before them. 

Ask anyone to conjure up a distant future where artificial intelligence has developed past the point of standard human capacities and you’re bound to presented with either a twinkling utopia, where automation has released the shackles of the working class and accelerates pursuits of culture and science; or a dismality where humanity’s own creations wreak havoc by outhinking us, outworking us, and outfighting us.

And some don’t think that future is so distant.

“The presidential candidate, Yang — uh — I can’t remember his first name,” said Fahy. “What is it?”

Her fumbled remark inadvertently highlighted the tech-entrepreneur-turned-minor-political-force’s underdog status in the presidential race while simultaneously undermining the surprise appeal of his platform, which is heavy on the doom-and-gloom consequences of unchecked automation.

“Andrew,” called out someone from the audience.

“Andrew Yang,” continued Fahy, “put out an op-ed about a week ago, where he cited the Council of Economic Advisors, that claimed 83-percent of jobs now earning less than $20-per-hour will have substantial parts of their jobs automated.”

Her point was to illustrate that popular discussion around automation and AI often contains points of alarm that turn people away from what she considers an opportunity for progress.

“I always say, ‘With challenges come tremendous, tremendous opportunities,’” Fahy said. “But government often struggles to keep pace with anything technological, so that’s why we have such an extraordinary panel here with us.”

The panel consisted of seven experts across a range of disciplines, and Congressman Paul Tonko, who was co-hosting and moderating the forum alongside Fahy. 

Despite the national framing, the relevance of the panel was focused on the Capital District, which is nestled within “Tech Valley” — a region of high technological investment that stretches from top to bottom along New York State’s eastern border. 

 The Capital District is home to the University of Albany, which offers undergraduate and graduate programs in computer science; a campus of SUNY Polytechnic Institute, where IBM recently opened an AI laboratory; and Rensselaer Polytechnic Institute, which itself hosts a business incubator and technology park. 

Many of the panelists had roots in these institutions: 

— Sanjay Goel is the UAlbany School of Business’s research director for Information Forensics and Assurance; 

— Deborah McGuinness is the chairwoman and professor of Computer and Cognitive Science at Rensselaer Polytech; 

— Shadi Shahedipour-Sandvik is the director of SUNY Poly’s nanoelectronics and technology; and

— Christopher Thorncroft is the interim director of UAlbany’s Atmospheric Science Research Center.

Additional to those panelists were:

— Andrew Kennedy, president of the Center for Economic Growth in Albany;

— Jim Malatras, president of SUNY Empire State College; and 

— Laura Schultz, director of fiscal analysis and senior economist at the Rockefeller Institute of Government. 

Each brought expertise in an area that, if not directly related to AI and automation, was or would be profoundly impacted by their development. 

Climate change

“We see weather disasters almost every day on the news,” said Thorncroft. “We have to be clever with the weather.”

He explained that both the prevention of further climate change and our defenses against it would benefit from an incorporation of AI.

Thorncroft said that, because decision-making in the wake of weather disasters is so critical, AI, which is essentially decision-making technology, is an obvious boon to human safety.

“AI is going to be very useful in tailoring forecasts,” Thorncroft said. 

And aside from making impending disasters known, better forecasts allow for more efficient harvesting of wind and solar energies, which Thorncroft termed “intermittent energies” that need to be accurately predicted in order to replace fossil fuels. 

In a vice versa, Shahedipour-Sandvik explained that fundamental to the “energy-hungry” AI field is the construction of energy-efficient hardware.

“Everybody, when they think of AI, thinks of the software aspect … but it’s really the hardware,” Shahedipour-Sandvik said. “[SUNY Poly] is focused on developing processes or hardware that are energy efficient. That requires a lot of innovation.”

Education

And where does that innovation come from? People. Of which there may not be enough, according to McGuinness.

“Data is exploding,” said McGuinness, referring to the fodder of AI systems, “but there aren’t enough data scientists.” 

“I think the biggest thing we have to do is look at education in the future,” said Malatras, who suggested that exposing kids to coding programs and teaching educational resilience would be imperative in keeping the U.S. ahead of countries like China, which already incorporate AI-based learning in their schools.

He also explained that learning can’t stop after high school or college, now that technology develops at a rate faster than ever.

“We’re not going to be able to stop AI,” said Malatras. “The question is: ‘How do we harness it?” 

Cybersecurity

If there were a doomsayer in the bunch, it was Goel, who was keenly focused on the increasing need for cybersecurity. 

“AI is going to be in everything we do for the next decade,” he said. “Today, people hack hospital systems and ask for money; tomorrow, they will hack into hospital systems and kill 15 people.”

Goel emphasized that today’s tech users are incredibly vulnerable.

“We keep falling for phishing,” he said, referencing a rudimentary and easy-to-avoid hacking technique that attempts to gain access to users’ private information by posing as a trustworthy source or company. 

“We need to invest — now — in very focused areas of [AI],” Goel said. 

Displacement

After the opening statements, Tonko posed his first question to the panelists: “Is there a way to take inventory of tasks that will be displaced by automation?”

McGuinness said it’s less about identifying specific jobs and tasks and more about figuring out the properties of those jobs and tasks.

“Someone said, ‘If it can be done in a second, it will be automated,” she said. “But we’re not going to replace the common sense that people have.”

Much of the panel agreed that jobs would not necessarily be lost to automation, but merely altered.

“When you go to the automated checkout at the grocery store, that’s one less employee working the register,” Schultz said. “But now that employee is walking the aisles and putting items into a basket for online shoppers.”

In total, Schultz said, about 53-percent of jobs contain tasks that will be automated at some point in the near future. 

“Translation has gotten pretty good,” McGuinness said later. “But people need to train and watch over these programs so the translations aren’t rote.

“I don’t think my colleagues’ [AI programs] are writing flowery poems,” she joked. 

Investment

Tonko’s second question sought guidance on how AI innovation can be advanced — the how and where of investment.

“I don’t think that’s something we’ve solved yet,” said Malatras. 

He said there are “areas and enclaves that do it well,” but that the various factors to contribute to those areas of boom won’t fit a national scale.

“It’s not something the government can do alone,” said Goel. “There needs to be a concertion of all parties involved.”

Government, educational institutions, and corporations would all need to devote time, energy, and money into figuring out what to do with new technologies.

The Capital District could become a leader in AI cybersecurity because the talent and programs are in place, Goel said, but he explained that less than 1 percent of AI funding is directed toward cybersecurity. 

Schultz said that the lack of money could be explained by the simple fact that not everyone benefits from expensive AI innovations, like those that threaten to cut investment payouts from fossil fuel.

“We need a carrot-and-stick approach,” agreed McGuiness. “We have a lot of sticks, but not a lot of carrots.”