Enlarge (credit: pitju / Adobe Stock)

This month, Google forced out a prominent AI ethics researcher after she voiced frustration with the company for making her withdraw a research paper. The paper pointed out the risks of language-processing artificial intelligence, the type used in Google Search and other text analysis products.

Among the risks is the large carbon footprint of developing this kind of AI technology. By some estimates, training an AI model generates as much carbon emissions as it takes to build and drive five cars over their lifetimes.

I am a researcher who studies and develops AI models, and I am all too familiar with the skyrocketing energy and financial costs of AI research. Why have AI models become so power hungry, and how are they different from traditional data center computation?

Enlarge (credit: Tracy J. Lee | Getty Images)

Michael Williams’ every move was being tracked without his knowledge—even before the fire. In August, Williams, an associate of R&B star and alleged rapist R. Kelly, allegedly used explosives to destroy a potential witness’s car. When police arrested Williams, the evidence cited in a Justice Department affidavit was drawn largely from his smartphone and online behavior: text messages to the victim, cell phone records, and his search history.

The investigators served Google a “keyword warrant,” asking the company to provide information on any user who had searched for the victim’s address around the time of the arson. Police narrowed the search, identified Williams, then filed another search warrant for two Google accounts linked to him. They found other searches: the “detonation properties” of diesel fuel, a list of countries that do not have extradition agreements with the US, and YouTube videos of R. Kelly’s alleged victims speaking to the press. Williams has pleaded not guilty.

Enlarge (credit: Flickr)

Early in 2020, I wrote about my experiences of moving to online learning: learning to use new tools, changing the way I taught, and dealing with the challenges of remote assessment. Sitting, unmentioned in the background, was the fact that the faculty where I teach had already agreed to revamp our electrical engineering curriculum, which directs the lives of students during three of their four years here. This raised a rather critical question: do we stick with the old in these trying times, or forge ahead with something new? In the end, we decided that we would press on with the new.

Then, just to add to the confusion, management decided that we should have two student intakes: one in August/September (the traditional starting time for new students), and one in February. This meant that, if we were fast enough, we could trial the new curriculum on a small group of students that started in February, rather than jumping in the deep end in September. After a lot of work, and with much material still to be developed, we think we are ready to roll.

One of the best parts of developing the new curriculum has been the criticism and feedback we’ve had from colleagues, students, and alumni. Now, with a month to go before we’ll be using the new curriculum, I want to open it up to critique by you, the Ars readers. I’m ready to be the Christmas roast.