+ THREE INSIGHTS FOR THE WEEK |
1. If you’re leading teams right now as artificial intelligence reshapes industries, you know that the biggest challenge isn’t implementing the technology itself. It’s developing a workforce that will ensure competitive advantage in this new era.
Here at MIT Sloan, our faculty members are studying how AI is transforming work in areas that matter to leaders — skills development, organizational design, and the jobs made resilient by uniquely human capabilities.
The five articles in this collection explore those capabilities and more. Inside you will find:
|
- Research examining why abilities like empathy, creativity, and judgment are becoming more valuable as AI advances.
- Strategies for balancing innovation and job quality as AI implementation continues to reshape work.
- Tips for preparing your employees to work smarter with generative AI.
-
A warning about relying on junior employees to teach new technologies to senior colleagues.
- Guidance on how to use AI to find and close skills gaps.
|
|
|
2. What if federal funding for biotech research had been 40% smaller over the past few decades? An “alternative history” reveals that breakthrough drugs for cancer and HIV might never have been developed.
MIT Sloan economists Pierre Azoulay and Danielle Li and co-authors analyzed 557 drugs approved by the Food and Drug Administration since 2000 and found that 59% of their patents cite research funded by the National Institutes of Health. Fifty-one percent cite studies that would have been eliminated under a 40% budget reduction — the level of cuts the current administration has proposed for the NIH.
Medications like Novartis’ cancer breakthrough Gleevec and Gilead’s HIV treatments Emtriva and Truvada all relied on research that would have been at risk, the report concluded.
“Massive cuts of the kind that are being contemplated right now could endanger the intellectual foundations of the drugs of tomorrow,” Azoulay told Fierce Biotech.
Azoulay and his collaborators had previously identified the time lag between when research is done and when benefits actually manifest. “Cuts today are going to have effects starting 15 years from now, roughly, and then accelerating from there,” he said.
|
|
|
-
Reducing operational carbon through hardware adjustments. Research from the MIT Lincoln Laboratory Supercomputing Center shows that “turning down” GPUs so they consume only about 30% of the energy they typically use has minimal impact on AI model performance while making hardware easier to cool. Stopping model training early can also save significant energy: About half the electricity used in training goes toward achieving the last 2 or 3 percentage points in accuracy.
- Leveraging algorithmic efficiency improvements. Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory are studying gains from new model architectures, compression techniques, and neural network pruning that solve problems faster while requiring fewer computing operations.
-
Maximizing energy savings through flexible scheduling. Engineers at the MIT Energy Initiative are building flexibility models that schedule computing operations for times when more electricity comes from renewable sources. They’re also exploring long-duration energy storage at data centers to reduce their reliance on fossil fuels during high-demand periods.
-
Addressing carbon embodied in data center construction. Building and retrofitting data centers — which require tons of steel, concrete, and cooling infrastructure — creates substantial carbon costs that are often overlooked in sustainability discussions.
|
|
| How to succeed with industrial AI
|
Applying systems dynamics principles to industrial AI can ensure faster and more significant business outcomes.
|
|
|
Despite widespread access to new technologies, many companies struggle or fail to integrate industrial AI into their manufacturing operations.
Companies need to harness data to build feedback loops that will lead to more intelligent insights, increased efficiency, and, ultimately, better decision-making, MIT Sloan senior lecturer John Carrier argues. Carrier explores this topic in the new MIT Sloan executive education course Strategy, Survival, and Success in the Age of Industrial AI.
Carrier’s action plan for industrial AI includes:
|
-
Gathering the right data and getting it to the right people. Data is central to industrial AI efforts, but low-quality data or an overabundance of ancillary data can do more harm than good. Make sure people doing the work have the information they need to make real-time decisions.
-
Fostering a culture that’s open to data-driven insights. The best leading indicators and AI-driven insights are useless if employees are distrustful of data or not open to information they may not want to hear.
- Remembering that simpler is better. The complexity of AI systems, models, and data can slow down results. Consider using smaller sets of models and data or limiting the scope of initiatives.
|
|
|
When managers use dynamic work design, they often feel, for the first time, that they are managing their organization rather than their organization is managing them.
|
– Nelson Repenning, MIT Sloan Professor, and Donald Kieffer, MIT Sloan Senior Lecturer, Co-Authors, “There’s Got to Be a Better Way”
|
|
|
|
One Main Street 9th Floor | Cambridge, MA 02139 US
|
|
|
|