Weather experts measure floods in decades, such as a 50- or 100-year flood. Computer science is seeing a 30-year flood of changes, thanks to artificial intelligence.
A speaker at Virginia Tech’s Day for Data last week ticked off a list of 1995 innovations that caused upheaval.
“Amazon sold their first book on the internet,” Ben Walker told a crowd of about 200 at the university’s Center for the Arts on Nov. 3. “There’s the Netscape IPO. EBay started. Craigslist started. Java came out. Windows 95 rolled out Internet Explorer. I mean, I can keep going.”
Many in the audience were not even born then, said Walker, a Virginia Tech graduate who is a partner at Alexandria-based management consulting firm Kearney. Regardless of age, the folks in the building on that Monday are working to reckon with the biggest changes to happen since that year.
“There was a massive, massive revolution for the internet and just the flow of information to the consumer in ways that it never has before, and we are experiencing that again this year,” he said. “It is remarkable to see just everything that is happening around AI and everything going into the industry, and I’m super excited about the potential it has and how fast everything has changed.”
Multiple speakers at the all-day event — including Mehul Sanghani, who grew up in Blacksburg and has become one of Virginia Tech’s most successful entrepreneurs — discussed the implications. A common theme emerged: As AI technology evolves, that 30-year storm might just become a recurring two-year storm with the same capacity to disrupt.
“I read … just about a week ago that the shelf life of overall skills has never been shorter, smaller, where 50% of the skills that you have today won’t be viable in two years,” said Sanghani, who founded Octo, a federal contractor specializing in information technology and digital services, and has remained its CEO since its 2022 sale to IBM for a reported $1.25 billion.
The onus is on the workforce to adjust to employers’ expectations, said Sanghani, who, with his wife, Hema, endowed the university’s Sanghani Center for Artificial Intelligence and Data Analytics, in Alexandria.
Here are some other takeaways from a conference that Jay Winkeler, director of Master of Science in Business Administration programs at the school’s Pamplin College of Business, told the audience would be “impossible to summarize.”
Who knows where it’s headed
While most agreed that changes would require responses, participants also acknowledged that they don’t know what those changes will be.
While AI has been around for decades, performing human-like tasks, generative AI is the revolutionary factor, in which machines can produce content based on materials they have been given as training. In popular use, it can be a souped-up — if unreliable — search engine that can create images, prose, poetry and music, while its professional potential includes complex code-writing.
In its very early stages, questions of trust and ethics have arisen, as the machines often “hallucinate” incorrect answers, and human creators have sued after alleging that tech companies have fed their work into large language models. Meanwhile, AI’s time efficiency and potential for prodigious output are putting people out of work.
The neural networks behind generative AI are so complex that even some experts say they don’t really know how they work. It is learning from itself, and that is giving it unpredictable qualities.
Panelist Rishi Jaitly, who founded and leads the university’s Institute for Leadership in Technology, told the audience that he recently brought in Vint Cerf, one of the acknowledged “fathers of the internet,” to speak to the institute’s fellows.
Responding to a question about what he does to stay fresh, Cerf told the fellows that he spends days trying to write the algorithms behind generative AI.
“He’s doing the math by hand to understand neural networks, and this era that has unfolded before his eyes,” Jaitly said.
Multiple panelists said that companies are hiring them as consultants to help them understand how to incorporate AI into their businesses. Leigh Sheldon, a Virginia Tech grad who works for McLean-based Guidehouse, said deploying AI isn’t a simple task.
“No one person out there has the crystal ball of what AI is going to look like in the future,” Sheldon said. “And I don’t know about y’all, but that, to me, is what keeps it incredibly fun in working life, because every day is continuous learning, and we’re doing that in support of our clients and growing throughout the course of our careers, which is super exciting.”
Job uncertainty and a possible leadership vacuum
While listening to the panelists made consulting sound like a good gig on the come up, AI’s speed and complexity are contributing to a lot of tech sector layoffs.
Megan Achinasi, a Virginia Tech grad and Norfolk Southern vice president, said that “unless AI was hallucinating,” it told her earlier that about two dozen companies over the past couple of weeks have announced about 100,000 U.S. job layoffs, about “a fourth of which were related to AI or restructuring due to automation and technology.”
She added, “Yes, I went to AI to get the answer for that.”
AI was not hallucinating that answer to Achinasi, who, along with Sheldon and others, was part of the day’s “AI and the Modern Workforce” panel. A report cited by several news sources this week said that technology firms have announced 141,159 job cuts in the year to date, an increase from the 120,470 over the previous 12 months.
“Technology continues to lead in private-sector job cuts as companies restructure amid AI integration, slower demand, and efficiency pressures,” according to a report by job coaching firm Challenger, Gray & Christmas. “In October, the sector announced 33,281 job cuts, up sharply from 5,639 in September.”
Government employment numbers remain unavailable due to the federal shutdown.
Plenty of Virginia Tech students were in attendance, many surely hopeful of work in computer science fields and at least some concerned about job opportunities. Sanghani said during the closing panel, “The Future of Learning: Leading Through Uncertainty in the Age of AI,” that he had chatted with some in the audience that day, hearing them rue the soft market of the moment.
“Folks are devaluing the entry-level folks,” he said. “But if you don’t have this sort of apprenticeship, who then serves eventually in the mid-career and leadership levels? If you’re not investing in that layer of leadership … you don’t have the creativity, you don’t have the leadership, you don’t have the judgment.”
But the gaps in AI are calling just for those things, which can’t be automated, he added.
Those who are interviewing for jobs might have a strange and incorrect idea about how to use AI.
Sheldon said that her company has observed some job-seekers using AI during their interviews.
“We can see their eyes following a prompt right as AI is generating responses to our interview questions,” she said.
Guidehouse has started implementing case studies and other challenges that preclude GPT help.
“So just keep in mind that while many employers are looking for and hiring AI and data scientists and analytic practitioners, using AI has a right time and place, just like anything else,” she said, noting that her company values unique and authentic people.
Trust is key
Those who use AI will outperform those who don’t, Kearney said during his morning session. It can reduce some task times from months to hours. That doesn’t mean that the technology is always a reliable partner.
Fact-checking is key, but having good data and doing the right things with it comes first, said Paul Needleman, a Virginia Tech graduate and data expert at software development company Snowflake.
Most important is keeping the information private. Sending certain data out to an AI model risks putting it in the public sphere, so spreadsheets with personally identifying information or credit card information can’t just go out to Chat GPT, Needleman said.
“You just now exposed your company to lawsuits,” said Needleman, who was on the panel with Sheldon, Achimasi and Amazon Web Services’ Keith Johnson.
Biases within models (both images and texts are susceptible, according to multiple studies) can lead to poor decisions that harm companies and clients. From there, fact-checking is essential.
“Who knows what this model is trained on? Is it hallucinating?” he said. “[Make] sure that you’ve dug into the details. You’re curious. You’re not just taking information that’s presented to you and regurgitating it back. Having a story, having an understanding of asking, why was this produced? How did it get produced? “
One word: Upskilling. Or is it reskilling?
One theme common to multiple sessions was continuing education. The words “upskilling” and “reskilling” came up numerous times.
Jaitly and Sanghani — on the education panel with Samiah Lattimore, APEX Center for Entrepreneurship’s executive director, and the moderator, Pamplin College of Business graduate programs director Jennifer Riley — discussed Apple University, the personal device giant’s training facility.
Octo also had a facility for upskilling, Sanghani said.
“From my perspective, I always espouse that, you know, career agility or adaptability is job security,” Sanghani said.
Adventure ahead
Needleman made a prediction about a new profession. While he didn’t give it a name, he described it as something that will emerge around semantic understanding of data and metadata — that is, to provide context and connection to the seemingly limitless data that companies can gather.
Keeping up with the latest tools, trends and emerging challenges can be a job in itself. A panel of more recent Virginia Tech grads talked about how they manage that information. Automated tools can send updates, or simply scrolling on sites such as LinkedIn can alert folks to seasonal software releases.
“If you’re reading the news and if you’re trying to set up more trends and new tools and everything, that’s great, but make sure you’re doing it in a way that feels easy, too, because we all work a lot,” said Grant Bommer, who works in finance for Leidos Holdings Inc. “Just find what works best for you, so it doesn’t feel like a chore. Because a lot of times this stuff can get monotonous, can feel like a chore, especially in a time … when these things change every week or month.
“You know, we look back where we were in January compared to now, and it’s a new world.”