At the MIT conference Generative AI’s ability to boost productivity, cut costs and spark new ventures was explored. But scaling and serious societal risks keeping implementations in check are also top-of-mind.
By Peter Krass
Judging from the recent BigAI@MIT conference, AI is at a crossroads.
On the one hand, AI progress is undeniable. One conference presenter described an AI system that can document common processes — such as paying an invoice — by simply watching a video of a person performing that task. Others spoke of impressive AI advances in healthcare, materials science, software development and more.
At the same time, many at The Business Implications of Generative AI @ MIT conference, held November 15 at MIT, expressed serious concerns over AI’s uncertainties, obstacles and risks.
Those roadblocks include training millions of workers with new AI skills, scaling AI pilots to enterprisewide implementations, correcting AI-generated misinformation and biases, and even the difficulty of identifying AI’s financial returns. Speakers also pondered complex economic and philosophical questions: Will widespread use lead to massive unemployment? Will AI l create even greater societal disparities?
“We’re at the ‘gradually, then suddenly’ stage of AI,” said panelist Rita McGrath, Academic Director for Executive Education at Columbia Business School. “The tipping point has happened….Now we’re busy sorting it out.”
McGrath, the author of a new book on permissionless organizations, also said that organizational issues, fairness and governance are top-of-mind for businesses. Beyond those, she added, “How do we make sure we don’t just unleash these weapons of mass delusion on the world?”
From left, Des Dearlove with panelists Scott Anthony, James Wilson and Rita McGrath.
HR, Marketing, Finance Impact
All speakers agreed that AI’s impact will be profound. H. James Wilson, Global Managing Director of Technology Research and Thought Leadership at Accenture, cited estimates of Generative AI having the potential to transform nearly 45% of all work done in the U.S. economy. “And if you drill down into some functions — like HR, marketing or finance — that number goes way up: 70% to 80% of the work in those functions is going to be impacted in a significant way,” he added.
For Wilson and other speakers, the goal is to bring humans and machines together in a workplace where humans complement machines and, in turn, machines unlock human productivity and ingenuity.
Data Matters
Some organizations are already charging ahead with AI and seeing big benefits. “Companies with GenAI exposure have been shown to achieve significantly higher market valuations,” said Sinan Aral, Director of the MIT Initiative on the Digital Economy (IDE), one of the BigAI@MIT conference hosts along with Thinkers50 and Accenture. “Those with significant data assets in combination with AI exposure are even more highly valued.”
Accenture is a prime example. The global professional-services company has some 50,000 data and AI practitioners, and it has invested $3 billion in their work.
“There have been so many amazing developments for GenAI technology and remarkable developments with all the foundation models,” said Lan Guan, Accenture’s Chief AI Officer and a panel moderator.
Another AI leader who shared lessons learned was Soumya Seetharam, Chief Digital & Information officer at Corning Inc. Her company, despite being a 173-year-old manufacturer of glass and life sciences products, has prepared more than 70% of its business-transaction data for use by AI. “It wasn’t easy,” Seetharam said. “We have 2,000-plus applications.”
Corning’s AI work is far from over. In addition to its structured AI-enabled data, Corning also has a huge volume of unstructured data — some of it a century old. “That’s the next frontier for us: business transaction data,” Seetharam told conference attendees. “How do we get this unstructured data into our platform, secure it, and make sure it’s available” to the business?
Tectonic Shift
Entrepreneurs, along with investors, have also rushed into the AI marketplace.
Lily Lyman, a General Partner at early-stage investor Underscore VC, cited estimates of $55 billion in venture funding invested industrywide to date, much of it directed to AI startups. Investors realize “how tectonic this shift is,” she explained. “They’re saying, ‘I want to be part of it.’ ”
From left, MIT’s Bill Aulet with panelists Heidi Messer, M.L. Carr, Stephanie MacConnell, and Lily Lyman discuss VC opportunities and entrepreneurship.
Some of those investors are likely to strike it rich, according to Bill Aulet, Managing Partner of the Martin Trust Center for MIT Entrepreneurship and a BigAI@MIT panel moderator. “While some parts of AI are over-hyped,” he said, “it’s going to be ubiquitous.”
To help would-be entrepreneurs, MIT recently introduced an AI-powered tool, the MIT AI JetPack. It takes would-be entrepreneurs through a 24-step framework to examine such factors as market segments, business models and pricing.
The BigAI@MIT conference also featured short presentations by 10 AI startups, all either funded by Milemark Capital — represented at the conference by founder Sebastian Barriga — or associated with MIT (or, in some cases, both).
Indeed, the businesses that presented are highly innovative and drawing interest. The video-watching AI provider, Klarity, has raised more than $90 million in financing, and its technology is now used by over half the 100 largest software companies.
Other startups presenting at the conference included:
- Catalan.ai: Developer of a solution offering Dynamic Pricing as a Service.
- Ikigai: Offers a GenAI solution for tabular data, using some 200 APIs and a no-code platform. The solution can be used for “what if?” scenario planning.
- Vertical Horizons: To meet AI’s nearly insatiable thirst for electricity, this company has developed a new type of power-supply unit. It promises to cut energy losses by a third and reduce demand for data-center rack space by half.
Gains and Pains
While recognizing AI’s enormous potential, many experts are troubled. “No one knows where this is going,” said M.L. Carr, a Partner at finance firm New Technology Ventures.
“We know [AI] is going to be huge….But the impact on people is a big part of this. We’ve got to figure that out.”
That attitude was echoed by Scott D. Anthony, a Clinical Professor of Business Administration at Dartmouth’s Tuck School of Business and author of an upcoming book on disruptive innovators.
“Yes, there are going to be great big productivity gains [with AI],” Anthony said. “But there are also going to be entire industries that are upended.” Offering an example, he said “advertising agencies are going to be really torn apart.”
University researchers have the job of looking to the big-picture, where AI can chart unlimited new paths. Sendhil Mullainathan, a Professor of Computer Science and Economics at MIT and a recipient of a MacArthur “Genius Grant,” asked attendees to think beyond current AI limitations. Instead of thinking about algorithms that do everything a human can, he said,
“What if we asked: What are things algorithms can do uniquely that humans could never even dream of? Those are the places where there are huge gains.”
Opinions differed on AI’s likely winners and losers. While Dartmouth’s Anthony sees “a democratization of technology,” Andrew McAfee, Co-Director of the MIT IDE, concluded that AI is likely to enlarge the already massive “competitiveness chasm” separating highly technical companies from others.
McAfee said that highly technical “geek” companies are prevailing. Their ranks include Netflix, Uber and Airbnb, many of which have higher market values than longtime industry leaders. “We haven’t seen the end of this story,” McAfee said. “I predict AI will widen the chasm even further. The world will be faster, but also more uncertain.”
The intersection of faster and more uncertain? That’s the new AI crossroads.
Do more:
- Register to watch videosof the BigAI@MIT presentations.
- Read more blogs on Medium hereand here.
- Browse the BigAI@MIT agenda
- Learn more about the MIT Initiative on the Digital Economy (IDE)
Peter Krass is a contributing writer and editor with the MIT IDE.