Science is a process for uncovering truth. More than ever, scientists must commit themselves to being stewards for this process and the imperfect forms of truth human cognition can uncover. AI has tremendous potential as a scientific tool, but tools need appropriate safety measures before deployment
And until I can have more confidence that AI have ethics and are grounded on truth, they'll be limited to very small, peripheral parts of my toolkit (formatting documents, generating draft infographics from specified data, etc.)
Until we can instill ethical standards and a better understanding of how to distinguish empty claims from actual scientific truth (which is really hard because you can't fully separate science from the assumptions and biases of scientists), stuff like this will keep happening
Expecting people to effectively debunk AI answers supercharges the problem of Brandolini's Law - that it takes far more effort to counter falsehoods than it does to create them, i.e. the bullshit asymmetry principle.
"But the humans are ultimately accountable for whatever they publish" say the AI apologists. Yes, but humans are not always ethical and responsible about these things, and we are fallible even when we are. AI decision-making systems are non-transparent but data intensive
Those LLMs and agents then go on to spread [mis/dis]information, and too many people subsequently expect those agents to do what they are nominally meant to do - give accurate answers based on massive amounts of data, e.g. the citation of junk studies in a peer-reviewed journal.
People wonder why I, and many other academics, are reluctant to embrace AI in our work. This is why. We lack tools to effectively instill ethics or high standards for certainty into LLMs and agents.
www.nature.com/articles/d41...
For those who are new to ILUC, there are better resources out there. I would humbly submit my YouTube video on the topic as one:
www.youtube.com/watch?v=eT06...
If there's interest in an ILUC basics webinar, let me know and that's absolutely something we can do.
Yes, though this is a very complex topic and we're intentionally framing it as a conversation among experts, aimed at an audience of experts and policymakers who work on the topic. So, I wouldn't recommend this for people new to the kinds of models or analytical techniques used in ILUC research
Please feel free to reach out if you have questions. To preemptively answer a common one:
Will there be a virtual option? Probably not, though we are exploring the possibility of recording some or all of the presentations and posting them online.
This is meant to be an opportunity to advance the discussion in this space beyond the "my model is better than your model" arguments that have dominated over the last few years, and help start researchers, like my team at the Biofuel Land Use Change program, developing more effective policy tools.
For any ILUC researchers, enthusiasts, or folks for whom those letters have become the bane of your existence: I cordially invite you to a research conference ITS-Davis is co-hosting with Prof. Aaron Smith of UC Berkeley, May 28-29 at UC Berkeley.
its.ucdavis.edu/events/revis...
Muscles... ligaments... same difference!
I'd like to announce an exciting new postdoc position we just opened up to work on new approaches to indirect land use change assessment and risk mitigation.
Please share through your networks!
its.ucdavis.edu/about-us/car...
Lovely new blog by @auffhammer.bsky.social, who clearly enjoyed a full helping of snark this Thanksgiving!
energyathaas.wordpress.com/2025/12/01/a...
I 100% agree, I have a Ph.D. in this stuff and I can't understand half of my bill (except the total I owe, they're great at making that clear)
Similar workshops are being held around the country, in case you're not in our neighborhood. See the link in the invitation to find one in your area.
Please share this announcement and I'm happy to answer questions!
Interested in the GREET model for transportation LCA? Have you found the GREET model to be very confusing and tough to work with?
We can help with that! ITS-Davis will host a 2-day R&D GREET model training workshop model in Sacramento this November, I'll be your instructor!
conta.cc/4lfRBoP
good morning, short thread on the raids in Camarillo, Ventura County, California. Friends of mine were tear gassed and fired upon with "less lethal" bullets for showing up with a bullhorn.
They carted away busloads of farmworkers.
Families torn apart. Children sobbing. This is Trump's America.
I just donated! RL Miller is the real deal. She shows up when things are tough.
True. The way to establish the water level a X year flood will hit is to watch that body of water over many 100's of years, or longer. You can't really do that while the climate is actively and significantly changing. We can model approximations, but our models are also in uncharted territory.
I have a new blog up at the UC Davis Institute of Transportation Studies: No, Your Gasoline Prices Will Not Rise by 65 cents per gallon in July. I'll give you three guesses to figure out what it's about.
its.ucdavis.edu/blog-post/no...
Right now, about 1/4 of LCFS revenue goes toward EVs; by the early 2030's it will be over 1/2.
I agree that capping the amount of lipid-based biofuel given credits under the LCFS would be good for the program and emissions, but killing the whole thing throws a lot of baby out with the bathwater
🌎 On #EarthDay, join us for our fourth annual #CAClimateSummit.
Register: theclimatecenter.org/summit2025
With policy experts, elected officials, business leaders, scientists, and activists, we'll explore how to draw down carbon and build resilience for a climate-safe future.
Lots of people saying, "just wait until the consequences hit!"
I would remind you that last time Trump was president, he botched a pandemic & a million Americans died. As a result, he was reelected & Americans turned on their own public health agencies.
Is there a ten cent thing that I can buy with a dime, but wouldn't buy if it would cost me a quarter? If so, then we should keep dimes, if not, then maybe we don't need cash smaller than a quarter. I can't think of that thing, however, a quarter is the smallest transaction we typically care about.
Agree that prices don't care about coins, but most transactions today are electronic, with cash's share likely to continue declining. So the question is what's the smallest unit of currency for which there is a significant need for precise cash transactions?
I'd submit that right now the quarter is the smallest meaningful unit of currency we have in the U.S. Anything smaller is kept for the sake of tradition or to let us complete the math without arbitrary rounding. Is that worth paying $billions to allow it to continue for cash transactions? I say no
If we can ignore the 1-2 cents the consumer or store might lose by omitting the penny, then why is the 2-3 cents they might lose without nickel a problem? Is five cents a functional unit of currency in the US? Do we buy anything for five cents, or is it just there to complete the transaction math?
It's hard to argue that the dime has much value as a mechanism of transaction these days, even harder to argue we'll still need it in 10 years. Do we need coinage smaller than a quarter?
Exactly. A problem which has the exact same solution: eliminate the nickel, too.
