Amazon taps U. of I. as artificial intelligence race heats up

The University of Illinois Urbana-Champaign’s connection with pioneering computer research dates back more than 70 years and is so famous it once found its way into film.

In Stanley Kubrick’s “2001: A Space Odyssey,” the sentient computer, HAL 9000, that torments the crew of a spaceship was based on technology developed at the university’s Coordinated Science Laboratory. The real-life lab became the proving ground for ILLIAC, a succession of supercomputers first built in1952.

Now, a new center for artificial intelligence research will launch Tuesday with the help of tech giant Amazon. The Amazon-Illinois Center for Conversational Experiences, or AICE, will work to make AI even more intelligent, according to UIUC professor and AICE Director Heng Ji.

Ji is also an Amazon scholar, a paid role where she works for Amazon once a week acting as an adviser or consultant for the company, pitching ideas and bridging connections between Amazon and universities.

The formation of AICE was first announced in March, and funding will also support Ph.D. fellowships in computer science. There will also be opportunities for faculty to submit research proposals for funding.

As the pace of AI research quickens and an increasing number of companies are turning to universities for their expertise, some critics are concerned with the ethical issues that partnerships between businesses and academia can create.

For example, the University of Toronto’s Faculty of Law school came under fire last month for not publicly disclosing a $450,000 gift from Amazon. The university returned the donation acknowledging in a statement “the important questions raised about the lack of full transparency pertaining to the gift, and the perception of external influence on our academic activities.”

M’hammed Abdous, vice president of teaching and learning at Old Dominion University, has been researching AI and its effects on higher education for the past year.

Abdous criticized the “fuzzy” objectives of the AICE and remarked on the privacy and security issues that can come with creating new open-source generative AI. Abdous also cited an apparent lack of input from the humanities side in AI models that involve language.

“One thing that AI is forcing us to do is really to increase the level of collaboration and cooperation. The people that have the ability with language are actually the humanities and social science side,” Abdous said.

Harley Johnson, associate dean for research at UIUC’s Grainger College of Engineering, said the benefits of the center — an outgrowth of the strong relationship the university already has with Amazon — flow both ways.

“Amazon gets access to expertise and it allows us as faculty researchers to translate our work into practice,” he said.

UIUC spokespersons and Ji declined to provide financial details of the partnership; the university has yet to respond to a public records request. Johnson said the center’s funding from Amazon comes in the form of both gift and grant funding.

Gaining intelligence

Ji said one of the center’s missions is to create something that does more than just “chat.”

“We want the system to have a sort of brain, a knowledge base … so that it can be more knowledgeable than a lot of humans, for example, and the conversation can be more informative,” Ji said.

Another goal for the center is to focus on furthering the ability of AI to connect with people and observe environments.

Ji described a system that could react like a human in everyday conversation, reading context clues and body language and even anticipating what someone might say next — even becoming a “good friend.” She gave the example of a machine that could move around and observe someone’s house, initializing and understanding the environment.

Ji said researchers at the center will have no commitment to create Amazon products, but its projects will align with goals Amazon has set out.

Aidan Sadovi/For the Sun-Times

UIUC is the latest university to partner with Amazon on AI, joining Johns Hopkins, Columbia University and Howard University. The growing number of partnerships speak to what one expert describes as an “AI arms race” between tech firms.

In January, Microsoft announced a $10 billion investment into OpenAI’s ChatGPT. A month later, Google invested $400 million in Anthropic, a rival to OpenAI. In April, Amazon launched Bedrock, a series of AI-related projects including materials that will help businesses and companies create their own chatbots with Amazon software.

Amazon CEO Andy Jassy has touted the company’s AI initiatives efforts. The CEO said in its latest earnings call that “every single one” of Amazon’s businesses were developing initiatives tied to generative AI, helping it become more cost-effective and improve the customer experience.

Robert Brunner, associate dean for innovation and chief disruption officer at UIUC’s Gies College of Business, said even though Amazon may look like they’re behind in the “AI arms race,” it would be wrong to rule them out.

“Every time you go to Amazon’s website and you shop or even just browse, they’re collecting data on you,” Brunner said. “They know who you are, and that’s how they get better and better and suggest things you might be interested in. They have a huge amount of data, which is a real strategic asset in the AI game and in the AI race.”