The Q Initiative

IBM’s Fletcher Previn ’00 talks quantum computing, rebuffing international hackers and human beings becoming subordinate bootloaders to a master digital race.

By Doug Daniels

o those of us who refer to our computer simply as “The Machine,” terms such as “artificial intelligence” and “quantum computing” can generate confusion, puzzled glares and can even conjure images of a dystopian future where humankind trembles in subservience to robot overlords. According to Fletcher Previn ’00, such fears are unfounded. Mostly.

Previn serves as chief information officer at IBM, a position that oversees a vast IT department with 12,000 employees and an annual budget of nearly $3 billion. That large team under Previn’s leadership is responsible for supporting 400,000 people throughout 170 countries around the world.

IBM operates the world’s largest privately owned network outside the U.S. government and is responsible for maintaining high-consequence mainframes for critical infrastructure related to clean water, the Department of Defense and the country’s health care systems, among other agencies. All this makes IBM and its clients prime targets for rogue hackers and global adversaries alike. A large portion of Previn’s time is devoted to securing those assets against hacking that originates in countries such as Russia, China, Iran and North Korea, especially with emerging technologies such as quantum computing—which dramatically increases processing power and has the potential to crack, in just seconds, encryption codes that would ordinarily take centuries to crack with traditional computers.

“There’s an alarming trend in technology where the bad guys seem to have the advantage,” Previn said, upon returning to campus in October to speak with students about his career and the challenges tech companies such as IBM are facing today and beyond the horizon.

While at Conn, Previn addressed the promise and the pitfalls of living in a world where so much digital data is produced and consumed every second of the day; he also discussed quantum computing and advanced artificial intelligence, areas that are increasingly intertwined and where IBM has been at the forefront.

“The challenge for me is to combat [the advantage of bad actors] by understanding IBM’s global footprint and networks and the systems that underlie those networks better than the people who are trying to break into them. At the end of the day, well-run networks really do make the life of our adversaries a lot more difficult,” he said, noting that IBM currently has a large quantum computer called the “Q System One” housed in its facility in Yorktown Heights, New York.

While there were early indications that Previn was interested in tech (for his 10th birthday he asked for a cake shaped like a computer), there was an assumption that he would follow in the footsteps of his parents—actress Mia Farrow and Academy Award-winning film composer André Previn—and pursue a career in entertainment.

“I spent a fair amount of time growing up around entertainment and on movie sets, and thought I would go into that line of work for a while,” Previn said. “While I was a student at Conn, I even worked as a writing intern at The David Letterman Show and The Conan O’Brien Show, but ultimately it became clear to me that I wanted to go into tech,” he said, explaining that he met his wife, Gillian Desjardins ’99, while working at the IT help desk in Conn’s Shain Library.

“We don’t want some kind of runaway optimization where we’re a subordinate bootloader to a master digital race.”

The Q Initiative

IBM has a storied history of innovation, from its intimate involvement with the Apollo moon missions to its pioneering developments in artificial intelligence dating back to the 1950s, when the company invented a program that could play checkers against a human opponent. The company’s contributions to medicine alone are too numerous to list, one of the most notable being the invention of LASIK eye surgery.

But today, the major trends shaping science, medicine and technology center on quantum computing and general AI. The IBM Q initiative experiments with using these advanced quantum systems in a multitude of fascinating contexts that will dramatically speed up disease diagnosis, detect fraud in real time and introduce ways to make energy use far more efficient. The applications for the massive jump forward in speed and processing power that are afforded by quantum computers also have the potential to revolutionize other spaces as well, ranging from the financial services industry to self-driving cars to pharmaceutical development and chemistry.

But as quantum computing becomes more accessible over the next decade (currently, operating the hardware requires a tremendous amount of physical space and energy, insulation from vibrations, and an ice-cold cooling chamber), hackers and nation-states with malign intent will also have more opportunities to meddle with financial markets, steal sensitive information and potentially wreak havoc on key infrastructure such as power plants. Organizations such as IBM are already anticipating these challenges several years out and are working to develop security systems. But with the enormous explosion of data swirling around, no human workforce can realistically keep up. That’s where artificial intelligence will play a key role.

Most people are familiar with basic examples of AI, such as voice commands for our mobile phones or Amazon’s Alexa, which can handle tasks such as playing music or reciting the weather forecast. That technology is fun and useful, but not at the level of the tech in which IBM is investing heavily, according to Previn.

In its infancy, IBM’s AI may have played checkers before graduating to chess, but the sophisticated versions that Previn sees as essential to how technology will operate in the years ahead will ultimately need to reflect not only the mental capacity of humans, but also the emotional intelligence that separates humanity from the machines.

“To truly be effective as we move toward the future of AI, one thing we’ll have to teach machines is empathy,” Previn said. “Without it, a worst-case scenario would be a machine concluding that there are just too many people on Earth and the planet’s not set up to sustain human beings. We don’t want some kind of runaway optimization where we’re a subordinate bootloader to a master digital race,” he added half-jokingly.

After pausing briefly to allow that terrifying prospect to resonate, Previn went on to explain the complexities of designing AI systems so they won’t include fundamental human biases inherited from unwitting programmers. The solution, he said, is to hold broad ethical discussions with input from a diverse spectrum of people from different cultures, races, genders and socioeconomic backgrounds—something the relatively homogenous tech world has had a tendency to struggle with.

“If we’re designing an AI system, there are choices that we have to make around ethics and morals and social norms,” Previn said. “A self-driving car has to make decisions around whether it should save the driver or avoid hitting a pedestrian.”

Other instances of bias that are far more nuanced could impact, for example, how banks decide to approve loans for people based on unfair risk assessments. “The data shows that somebody with a family member who has been incarcerated is five times more likely to default on a loan. From a bank’s perspective, it would be logical to build an AI system that can pull that information from a data set and deny a loan automatically. But is that ethical?” he asked.

Previn said he expects quantum computers to be commercially available within 10 years and believes they’ll mark a historic shift in technology. New drugs and financial models discovered in seconds, security challenges large and small, and robots that can learn and think the same way humans do are only part of the equation.

Quantum computing may well set off a new international, tech-based arms race of sorts among competing governments and economies.

“There’s certainly a lot of interest from the United States government in doing what’s necessary to win the quantum race,” he said. “And there’s really no time to waste, because unlike traditional computing—where one person’s got the fastest supercomputer but it doesn’t really matter because the second fastest is still pretty close—being only slightly better at quantum offers a huge advantage and is a big problem for the adversaries, because competing with and defending against that power is impossible.”