Tia Ghose

Micky Tripathi is the top official in the U.S. Department of Health and Human Services spearheading much of President Biden’s agenda for regulating how artificial intelligence is used in healthcare. Mohana Ravindranath takes a closer look at the challenges ahead of Tripathi, especially at a time when the U.S. is already lagging behind Europe in regulating AI. As the HHS assistant secretary for technology policy, Tripathi’s success, or failure, could have a huge impact on how this technology with transformative potential is adopted across the health care system, especially given its propensity to hallucinate and, if deployed irresponsibly, exacerbate historical biases. He is in the hot seat at a time when health systems and insurers are racing to deploy AI-based tools to reduce administrative burden and maximize profits. And, as investors pour billions of dollars into AI ventures, industry groups are begging for rules that don’t stifle innovation and let clinicians and patients benefit from AI’s potential to advance medicine.

8

Here's a writing tip for researchers, academics, and others who suffer from Sententia Longitis. Find a friend who has diminished lung capacity. (I'm sure you know an asthmatic or a chain smoker with COPD.) Get them to read your writings out loud. If they can't finish your sentences in one breath, your sentences are too long. Split them, snip them, or just nip them. (Your sentences, not your friend.) Then thank your friend and wish them good health. #writing #researchwriting #language #funny

38

10 Comments

Many writers and editors think the Associated Press Stylebook is against using the serial (Oxford) comma. That’s not true. AP style *does* recommend using the serial comma in complex lists. For example, if one of the items in a list contains a conjunction, AP style calls for a serial comma. Like this: Fritz enjoys snacking on bananas, cheese and crackers, and ice cream. Additionally, if the items in the list are long, AP style calls for a serial comma. Like this: Helen changed the car’s worn tire before embarking on her big road trip, Raul shined the fenders until he could see his reflection in them, and Sara double-checked all fluid levels. AP does not recommend using a serial comma in a simple list: Dante bought strawberries, sliced turkey and rye bread. This is in contrast to most other style guides, including The Chicago Manual of Style, which calls for a serial comma in all lists. To learn more differences between AP style and Chicago style, check out Dragonfly Editorial’s field guide on this very topic. It’s titled “AP Style and Chicago Style: 15 Major Differences,” and the link is in the comments.

681

72 Comments

Generative AI has emerged as a promising solution to the U.S.’s broken healthcare system. Its potential applications include everything from assisting clinicians with diagnosis to creating comprehensive treatment plans and even administering therapeutic interventions. Large language models (LLMs) have demonstrated capabilities that can oftentimes parallel human performance in these categories, providing accurate instructions and data. #AI #GenAI #Healthcare

5

In case you need to hear this: keep your headlines short!   I frequently find myself at a tug-of-war between being comprehensive versus concise in my news headlines. But according to a recent research paper in Science Advances, general audiences are drawn to simple headlines more than complex ones: https://lnkd.in/gxsVKbPt Yet what intrigued me the most about this paper's findings was how journalists/writers sway the opposite: they tend to prefer complex headlines. This suggests that "those who write the news read it differently from those who merely consume it." As an early-career communications professional, I find this research a great reminder to consistently ask "Who's the audience?" and check thoroughness with approachability. 

15

1 Comment

It’s great to see this STM report on Trusted Identity. I’m a firm believer in eventually pushing the research integrity question all the way up the chain to the point of submission. It’s fascinating to observe two parallel tracks of change in scholarly user authentication. On one hand, efforts are focused on transforming reader authentication, driven in part by browser privacy changes. On the other, there’s a growing emphasis on reimagining author authentication, driven by goals of integrity. Of course both readers and authors are the same people and so we have the vision of a single new system that covers both, delivers that much needed trust, and enhances audience understanding for stakeholders across the ecosystem.

6

For Fast Company I looked at Akido Labs, a unique healthcare and tech company trying to leverage AI to help doctors, including those service Southern California’s homeless population. The potential—and potential pitfalls—of using artificial intelligence in the medical field https://lnkd.in/gN9aYwEu

24

1 Comment

Editing pet peeve: the goofy obsession with the Oxford comma. Oxford enthusiasts—whom I often refer to as "Oxfordnistas"—insist that the serial comma is the ultimate arbiter of clarity and slayer of ambiguity. And they demonstrate this overstated power through a series of reductionist—albeit humorous—memes. Below on the left is one such example. Again, this meme assuredly suggests a moment where a lack of a serial comma may confuse readers. Rather than express a list of the three, the Oxford-less sentence seems to imply an appositive suggesting the exotic dancers are, in fact, Stalin and JFK. But what happens if you change strippers (plural) to stripper (singular). One good contrived strawman argument deserves another, right? On the right is the counterpoint. Here, the sentence with the Oxford comma introduces ambiguity. Is that JFK on stage three collecting singles, or is somebody else? The trick is simple: When in doubt, rewrite. "We invited JFK, Stalin and the strippers." See how this is clearly a list of three? See how I didn't need the comma to do so? Ultimately, the Oxford comma is a stylistic convention, so it is a matter of personal preference. You can also write, "We invited JFK, Stalin, and the strippers." That *also* works. Either way, stop pretending that your prescriptivist take on the Oxford comma is holy writ. #Oxfordcomma

2

1 Comment

I have a new story live in Grow by Ginkgo! Read on to learn how emergent scientists are using LLM architecture for the greater good 🤖🔬🚰 https://lnkd.in/ekmMib3D

14

In conversation with Carl D. Marci, M.D. from OM1, Inc. regarding the future of machine learning, large datasets and its potential applications in earlier detection of Alzheimer's disease:

15

When I worked in science, I felt the strain of needing positive results to publish when most of science produces negative results. It’s exciting to see someone value negative results enough to create a pathway to publication. Negative results provide important information for scientists to use as they evaluate their experimental design and focus. Interestingly, AI has been a pivotal motivator for this push to share what did not work. “We’re interested in stuff that was done methodologically soundly, but still yielded a result that was unexpected.,” says Editor-in-chief Sarahanne Field at the Journal of Trial & Error. To read more, you can go to Nature’s article on the topic: https://lnkd.in/e_ZueGed #science #negativeresults #publishing

1

'Imagine navigating a busy freeway blindfolded. That's what predicting public health risks without big data feels like, as management consultant Geoffrey Moore aptly put it: "Without big data, you are blind and deaf and in the middle of a freeway." This analogy perfectly captures the essence of why open source data is indispensable in improving health outcomes and interventions at the population level.' I really found it interesting writing this piece for EPCON on how open source data can be used to improve public health decision making. https://lnkd.in/dikW7H_a #PublicHealth #OpenSourceData #HealthcareInnovation #DataAnalytics #HealthTech

4

1 Comment

This week on Scientific American's Science Quickly news roundup: AI Is Getting Creepier (read: learning how to neg you) and Risky Cheese Is Getting Trendier #podcasting #sciencepodcast #learning #AI #OpenAI #BirdFlu #health #audio #news #science #sciencenews

3

Part 2 in discussion of Associated Press Stylebook's "obesity" entry. (see part 1 at https://lnkd.in/gFn8QyF9) And AP does not take care in discussing body mass index (BMI). It says simply, “Many experts acknowledge that BMI is a flawed metric because it does not directly measure body fat, but say it remains a useful and convenient guide.” That explanation does not encapsulate why BMI is a flawed tool (https://lnkd.in/ebaw3DQm). Not only is it bad at what the health establishment uses it to do, but it has a racist history. As professor Sabrina Strings says, “Obesity science has always been a (racist) form of pseudoscience that relies on statistical correlations based on a limited portion of humanity" (https://lnkd.in/eTymQTva). AP handles so many topics admirably and with nuance–like discussing the problems of bias in AI systems in its artificial intelligence entry. I wish it took the same care to explain why relying on BMI (and therefore the categories of “obese” and “overweight” altogether) is problematic. Similarly, explaining who and for what BMI remains useful and convenient would be helpful. Overall, I recommend AP revise this section to provide more context and more appropriate guidance for journalists trying to cover higher-weight people. This is a pressing need, especially given the increasing coverage of drugs like Wegovy. In the meantime, writers and editors should look to language preferences from the people they’re writing about. For example, a report by NAAFA - National Association to Advance Fat Acceptance and researcher Pamela Mejia on fat bias in the news has some recommendations for journalists, one of which is “Avoid the ‘o’ words (obesity and overweight) and include other ways to talk about fat people.”  https://lnkd.in/ei37A-Ru For people who are interested in learning more about this kind of inclusive language, I recommend the ACS Inclusivity Style Guide’s section on body size (https://lnkd.in/eQN5fFG2).

3

I think this will be a net positive for Gannett's journalists, frankly. Content creation is hard - figuring out what to report, navigating access and recruiting sources, collecting facts and quotes, analyzing it all, and crafting the article/photo gallery/video. It gets a lot harder when you *also* have to create the presentation and troubleshoot any problems in publishing that presentation (which used to be a separate job, **cough** mine!). I would have loved the help of a large language model to draft a good headline or summary "deck". https://lnkd.in/e3uBuJWi

7

1 Comment