AI initiative

WPI has had AI research and courses since 1985, but with the major advances of the last few years the need for a specific degree program became clear. Going beyond computer science, elements of AI are now being used in health care, finance, manufacturing and space exploration.

The pace of advancement is only going to increase, said King.

“Change won’t be over 50 years but 50 days or minutes. Everything is moving at the speed of light.”

Back in 2020, the Future of Jobs Report from the World Economic Forum predicted that AI would potentially replace as many as 85 million jobs by 2025, she said, while generating 97 million new ones.

Building a degree on high speed

Though conversations about a new degree program began in 2019, the intensive design work began last August, said Elke Rundensteiner, professor of computer science at WPI and one of the program’s primary architects.

“The speed is unusual,” she said, as the process can sometimes take years on some campuses.

Rundensteiner and her team were able to draw on existing AI course offerings as well as her previous experience in founding the school’s data science program, which had included related aspects.

“The data science program uses many machine learning courses in it so doing this AI degree came very naturally,” she said. Machine learning is a subfield of AI in writing computer programs that are able to learn from experience.

New courses were added based on consulting with an advisory board of industry representatives, who described the skills they were looking for in new hires. “Advisory board members keep on telling me we need employees who can not just create but maintain these AI models – fix them, adjust them,” said Rundensteiner as an example. This request led to the creation of the machine learning operations course.

Some were requested by students, such as generative AI, which includes ChatGPT, the program through which the wider public was introduced to AI, which took the internet by storm, making AI a household term.

The degree program is built around the principle of required courses covering core concepts, after which students will be able to choose their own industry direction.

“Because AI is going to be pervasive, we felt it was the right thing to say, ‘Here’s a core that everybody needs to know,’ then students can pursue any aspects of AI that makes sense to them,” said professor John McNeill, WPI’s dean of engineering.

“Rather than just lay out a 10-course checklist and say here’s your degree, we have a more flexible path.”

Among these core areas, which include machine learning and data modeling, is responsible and ethical use, she said, where students will learn more about the policies they have to follow and the harm they could do.

Ethical use

King believes WPI is uniquely suited to walking this line between the technical and ethical implications, as the intersection of technology and society is studied closely through the Department of Social Science and Policy Studies.

All students are required to take courses in the arts and humanities, integrated alongside the STEM curriculum.

“We don’t just teach (students) how to use the tools but how to use them with intent and to understand the social and cultural ramifications,” she said.

This will be essential when studying AI, said Rundensteiner, due to the field’s sheer pervasiveness.

“It’s now being used in all facets of our lives,” she said. “(AI systems) learn only from the data that is put into them, which is shaped by our environment and society,” she said.

Which means that a machine can and will incorporate human bias, she said, depending on who the data comes from and some of these machines are being used to make decisions that affect large numbers of people – “who gets a scholarship, who gets a job, who gets ahead” – which places the responsibility of understanding what AI can and can’t do on those deploying the system.

“If AI is going to be touching people’s lives, they need to understand how it works and have some healthy skepticism,” said McNeill. “Not just thinking if the AI says it then it must be right.”

King is aware of what bias inherent in data could mean for diversity or simply equal opportunities. “As a woman of color, I’m always interested that we have more people at the table and make sure that tech is not just for certain people,” she said.

Though as a self-styled relentless optimist, she believes that this will be addressed through training a diverse workforce that will be the ones inputting this data.

“AI can be used against minorities but my approach to have them in the room to begin with,” she said.

Courses covering the ethical aspects of AI are not new, said Carolina Ruiz, professor of computer science, who was part of the design team, but with the new program will now “be more explicit.”

“No matter how good a technology is, there is the potential for people to use it in a negative way so as a society, we should regulate and prevent that,” she said.

Despite this, Ruiz believes that the public has been inundated with the negative implications of AI and not enough with the transformative benefits, she said.

Experiencing a revolution

“We’re at a pivotal point, because AI will stay with us for the rest of our history, making our lives more productive,” said Ruiz.

The age of the internet will further amplify the effect of new technologies and advances by greatly enhancing the speed with which more people can reap the benefits, Ruiz said.

“Before, if you develop a new car, things are not necessarily easy to move around the world,” she said.

“With software, we can bring these tools to people in remote areas in a matter of seconds.”

Most of the doubts surrounding revolutionary technology often center around what this means for jobs. “There will be volatility in the job market but so much more of an upside, once people get the education and over the fear of something new,” said McNeill.

Just some fear their jobs will be taken by AI, McNeill recalled when computer-assisted design for circuitry manufacturing became prevalent in the 1980s, and many in the industry thought it would replace designer engineers.

Instead it offloaded basic tasks to computers and allowed the designers to make better products and even created new jobs.

“AI will be the same way,” he said. “Some of those old drudgery jobs went away but a computer can’t give all the answers and will still need a human being to be creative.”

This example demonstrates where AI can shine in the workplace, said King.

“AI should be doing the tasks that are rote – computers are great at that,” she said. “They don’t get tired of doing the same thing over and over but humans do. What AI can produce for us is the ability to be more creative.”

Creativity is why a machine can never completely replace a human, said McNeill.

“The AI tool will get you 90% of the way there – you still need 10% of human input to do something genuinely novel that is going to have value.”