Musings and Writings

Some articles and papers that I enjoyed writing. Any typos or factual inaccuracies are, as always, the author's sole responsibility (i.e., me).

That said, always happy to kick the tires on these topics and argue and maybe come up with an interesting research project...why not?

The Impact of Generative Artificial Intelligence on Socioeconomic Inequalities and Policy Making

June 2024, PNAS Nexus 3(6), with Valerio Capraro, Daron Acemoglu, et al.

This perspectives piece covers a wide range of applications and implications of new generative AI technologies.  We investigate the effects of these tools on the workforce, education, healthcare, information, and policy — proposing a variety of policy directions and areas for further research to ensure that the future of generative AI development is pro-worker, pro-human, and leads to shared economic prosperity.

Available from Oxford Academic: https://doi-org.libproxy.mit.edu/10.1093/pnasnexus/pgae191.

Capraro, Lentsch, Acemolgu, et al (2024-06-11).pdf

From Automation to Augmentation: Redefining Engineering Design and Manufacturing in the Age of NextGen-AI

March 2024, MIT Press, with Md Ferdous Alam, Nomi Yu, et al.

This paper is one of a series of AI impact papers funded by MIT to articulate effective roadmaps, policy recommendations, and calls for action across the broad domain of generative AI and its effects on society.  

We focused on investigating the usefulness (or lack thereof) of generative AI in the domain(s) of manufacturing and engineering design.  Pulling from a number of interviews with industry experts, researchers, and practitioners, we conclude that today's generative AI tools are simply not usable in engineering or manufacturing. 

This can be attributed to several key deficiencies, including: the inability to provide robust, reliable, and replicable output; lack of relevant domain knowledge; unawareness of industry-standards requirements for product quality; failure to integrate seamlessly with existing workflow; and inability to simultaneously interpret data from different sources and formats. 

We then propose a framework to develop the next generation of Gen-AI tools for design and manufacturing: "NextGen-AI"

MIT Press, from An MIT Exploration of Generative AI: https://mit-genai.pubpub.org/pub/9s6690gd/release/2.

Alam, Lentsch, Yu, et al (2024-03-27).pdf

The Hollywood Writers' AI Fight Is Everyone's Fight 

August 2023, Project Syndicate, with Daron Acemoglu and Simon Johnson

We argue that, if properly designed and implemented, AI could not only be a very powerful tool to boost human creativity — it would also be a source of shared prosperity. Based on our read of economic history, however, this will not happen unless workers have a say in how these new AI tools are used.

If workers don't have a say — if we focus on short-term cost-cutting over long-term planning to develop talent and maximize productivity — then we risk "over-automating" work and putting too many workers out of jobs.

This is definitely bad for workers, but ironically it may actually be even worse for firms. If we haphazardly kill the talent pipeline in knowledge-work industries (a central concern for the striking Hollywood writers and actors), who is left in 30 years to come up with new ideas? Gen AI is very powerful, but in its current form it is not as "creative" as it is "automative."

Fun fact: this is my first byline since being a feature-writer for the T.F. Riggs High School paper, breaking hard-hitting news like the introduction of pumpkin to Starbucks' PSL in 2015 (not kidding).