AI Goes to Work
At two MIT CIO Symposium panels, experts describe present and future work-worlds where AI is the engine.
By Paula Klein
Is the future of work already here? And if so, how do we prepare for it?
Those were among the key questions addressed at the recent MIT CIO Symposium. During a discussion, The Future of Work in a World of AI, ML, and Automation, led by IDE Director, Erik Brynjolfsson, academics said that digital technologies can offer more opportunities than many people realize.
For example, Jason Jackson, MIT Assistant Professor in Political Economy and Urban Planning, noted that employment platforms help match jobs and people; for transportation, autonomous vehicles, and “urban mobility” apps like Uber, can help people get around cities more efficiently, and AI can cut energy costs and optimize workloads at large data centers. At the same time, he said, the disruptive nature of these on-demand models are creating “precarity” for long-term workers who must become contractors and entrepreneurs with “no guarantees” of permanent work, benefits, or wages.
New skills are needed, said panelists including Iyad Rahwan, Associate Professor of Media Arts & Sciences, MIT Media Lab, and Elisabeth Reynolds, Executive Director, Work of the Future Task Force. Reynolds said only about the 5% of jobs are replaced by technology compared with many more that are enhanced by it.
Watch the full video of the discussion here.
Another panel looked at the future of work with a different lens — current AI business implementation tactics and strategies. MIT IDE Fellow, Michael Schrage, asked IT leaders from DBS Bank, Adobe, and KAYAK what skills they look for in AI talent and what metrics they use to measure AI business improvements today.
David Gledhill, Group CIO at DBS, said his global bank is starting from the top — training the top 250 executives to understand AI and what it can — and can’t — accomplish. “Leaders have to be conversant” and ask the right questions, he said. That eliminates the “magic” and hype surrounding AI.
KAYAK may have an advantage as a “born-digital” company, but CTO Giorgos Zacharia, said AI isn’t always the right solution for a problem. “Tools have to drive outcomes and customer satisfaction,” he said. AI drives much of what KAYAK does behind the scenes, but simply completing a travel transaction on the site is a metric for success, Zacharia said. When too many changes are made to the user interface, that rate can fall.
At Adobe, CIO Cynthia Stoddard, also said that prioritizing external customer experience drives most internal AI design efforts.
Executives also recommend these key implementation points:
- Design and operate with a customer-first mindset. Will AI offer more speed, ease, and insights?
- Use analytics to anticipate and head off problems; remove friction.
- Emphasize and require high data quality.
- Put machine learning to use solving problems and test, then scale.
Watch the video of the session here.
