AI on the Lot the 4th Edition
...
This year the conference attracted 2,460 people to Culver City on May 27 and May 28. For those two days nowhere on the planet had a denser concentration of top talent working in the AI media space than a two-block radius of Culver City. ...
In 2026, something clearly shifted with AI video and Hollywood. And this May, AI on the Lot served as a kind of coming out party for this sizable vibe shift. ...
Micro-dramas, playable IP, character avatars are versions of what many people are calling neumorphic forms of media. These forms, while nascent, represent where cultural energy, consumer attention, and investor money are going. ...
... we made a small task force to give out free tickets to artists. We’ve done this every year, largely through our own networks. The conference size (2500 attendees) meant we needed to do this at scale. Simply reaching out to friends no longer worked. The solution we found was to identify interesting artists and give them bundle of free tickets to give their artist friends. We had two rules: the ticket must go to someone who is a working artist and the person must not normally go to conferences. This strategy worked well to bring interesting people to the conference and counteract any danger of it feeling too corporate. ...
See the full story here: https://www.linkedin.com/pulse/ai-lot-4th-edition-mike-gioia-ixanc/
When AI art has no author: MIT study finds generated images often can’t be traced to any training data
PhilNote: we're back to "correlation versus causation"
... at sufficiently large scales, they find, you can often remove any single image from the training data, or every image by a given artist, or every photograph of a given person, and the generated sample doesn't change.
And if removing something changes nothing, the researchers argue, it can't be said to be responsible for anything. "If you take away a piece of data and the output of the model doesn’t change, then that piece of data didn’t affect the output," says Zheng Dai SM ‘21, PhD ‘24, former MIT CSAIL researcher and lead author on the work. ...
"All previous methods were approximate," says MIT professor and MIT CSAIL principal investigator David Gifford. "They really could not absolutely show that deleting individual things did not change the output. This paper introduces the first method that is absolute. You're actually deleting the inputs and deleting all influences of the inputs. This is the first exact method for doing large-scale deletion efficiently and showing that the results don't change." ...
Exploring a counterfactual universe
With ablation working, the researchers could finally ask their question at scale. Take one generated image, then imagine every alternate version of it, each produced by removing a different piece of the training data. The team calls this the image's counterfactual universe. The distance between the original and its most different alternate, the counterfactual radius, captures the most any single piece of training data could have mattered. ...
The privacy paradox
The implications run in a direction that surprised the researchers themselves.
Gifford sees the finding as bearing directly on the legal question of whether model outputs are derivative works. “One way to think about this is that these models are creative. They are not simply copying what they are fed, but creating brand new outputs. If those outputs have nothing to do with any individual piece of training data, that raises questions about fair use, about whether the outputs are themselves copyrightable as novel works, and about how authors get compensated when what comes out of a model isn't attributable to anything on the internet.” ...
...this paper provides reason to think that attribution will fail for interesting models. Instead, technologists and courts will need to resort to other methods for assessing copying.” ...
See the full story here: https://www.csail.mit.edu/news/when-ai-art-has-no-author-mit-study-finds-generated-images-often-cant-be-traced-any-training
Higher education is trying to stay relevant with AI. Some students don’t want it
...
This year, more than ever, Watson said universities and colleges are grappling with the unethical murkiness of AI — from environmental impacts to data centers to intellectual theft concerns and the potential effect on jobs — while also trying to stay relevant in the current marketplace.
Young adults in the U.S. are increasingly wary of AI and its impacts, according to a recent Pew Research poll. At the same time, 91% of employers think it is very or somewhat important that college graduates gain skills in college to use AI, according to a recent AAC&U report.
Meanwhile, institutions are working to attract students amid a shrinking pool of traditionally college-aged Americans and widespread concerns about the cost and value of a college degree. ...
While he acknowledges student concerns about the co-major, he repeatedly emphasized that it is optional. ...
Bryan Alexander, senior scholar at Georgetown University, said AI is the most divisive technology in higher education. He said the backlash at Suffolk against a major, rather than AI in general, feels “groundbreaking” and like something to watch. ...
See the full story here; https://www.masslive.com/education/2026/08/higher-education-is-trying-to-stay-relevant-with-ai-some-students-dont-want-it.html
OpenAI wants California to strengthen its newly passed AI safety law
PhilNote: this sounds great until you think about how it would actually work. Will they actually hire experts to monitor and enforce independently from the companies driving the work, and will they fund the resources needed to do effective testing and verification?
...
In a post from OpenAI's Global Affairs team on LinkedIn, the company laid out its case for amending SB 53, the Transparency in Frontier Artificial Intelligence Act, which Governor Gavin Newsom signed into law back in Sept. 2025. OpenAI says the law needs new teeth, specifically, requirements that frontier AI models be monitored during training or evaluation for signs that they could bypass a third party's security controls or obtain confidential information they shouldn't have. ...
It also created CalCompute, a public computing consortium meant to give safety and equity research a boost, and it requires the state to revisit the law annually as the tech beneath it shifts. Newsom pitched it at the time as proof that California could protect the public without choking off the AI industry that has become such a huge part of the state's economy. It was also called a stopgap for Washington's inability to pass anything resembling a real federal AI policy. ...
See the full story here: https://mashable.com/tech/openai-urges-stregthening-of-california-ai-safety-bill
We still don’t know how people are really using AI
... It’s a public platform that aggregated and analyzed real AI conversations with popular models like Claude and Gemini that were collected with users’ consent through seven existing datasets. The intent is to provide independent sources of information that can help researchers and policymakers assess how people are using generative AI. ...
When the AI Observatory researchers applied Anthropic’s methods to their dataset, they found that nearly half the conversations—48%—would have been filtered out. Those non-work-related conversations were more likely to involve health and relationships (44.2% versus 31.2% in Anthropic’s analysis), adult or illicit topics (7.9% versus 2.1%), harassment and hate (27.5% versus 5.66%), and sexual content (16.7% versus 2.4%). (OpenAI’s 2025 report on ChatGPT, similarly, found that only 30% of consumer use was related to work.) ...
The fact that the AI Observatory’s dataset draws from voluntarily provided sources means it’s probably underrepresenting sensitive uses, which people may be less likely to share. Thus, the researchers caution that its findings are not indicative of all AI use. ...
See the full story here: https://www.technologyreview.com/2026/08/18/1142226/how-people-use-ai/
Minnesota defends AI ‘nudification’ ban after lawsuit from Musk’s xAI
Minnesota Attorney General Keith Ellison said Elon Musk's artificial intelligence company xAI should not be allowed to halt the state's ban on AI "nudification," telling a judge xAI's lawsuit challenging the ban is misguided and likely to fail. ...
The law, which took effect on August 1, prohibits website operators, software developers and others from allowing users to create realistic images depicting an intimate body part that is not shown in an original photo of an identifiable individual. ...
xAI has begun suing users whom it alleges are evading Grok’s technological blockers to create sexual images of people without their consent.
In its lawsuit, xAI contends Minnesota law is overly broad and contains “no exception for images of artistic, scientific, political, satirical, educational, medical, or religious significance.” ...
See the full story here: https://www.reuters.com/legal/government/minnesota-defends-ai-nudification-ban-after-lawsuit-musks-xai-2026-08-17/
Vote 16 Culver City
I am working on a campaign against Culver City Measure YV - Vote 16 Culver City. The measure would lower the voting age on all Culver City elections to age 16. Mayor Freddy Puza cast the deciding vote in favor of putting it on the ballot and coauthored the Argument For Culver City Measure YV - Vote 16 Culver City. I am helping write the Rebuttal to the Argument For... You can learn more about this issue at www.Vote16CulverCity.com , a website I am very proud of building!
AI agents aren’t legally responsible for any harm that they cause, experts say. So who is?
...
The agent was being helpful, he wrote, but his experience showed that if you gave an AI permission to do something for you, it would “often discover paths you did not explicitly ask it to look for”.
...Australian law applies only to people, not virtual beings. Experts say the person - or business - that deploys the AI agent is legally responsible. And they warn that those “deployers” often have very little idea about their legal liability. ...
“[What if] it engages in racist, sexist, misogynistic language?” Paterson asks. “Then you might be asking where the guardrails are that the developer has provided. You should be putting out a product that is reasonably safe.”
In that case, the developer could also be held responsible for not putting basic guardrails in place, which is where some legal ambiguity could creep in, she says.
Paterson said the law is different to, but informed by, ethics. ...
“As soon as we allow AI agents to act for us, they’re acting on the goal we give them, and if we don’t give them a whole bunch of parameters, the agent’s just going to try to achieve that goal [in any way],” Johnson says. ...
“I hear a lot of people saying ‘Oh, I made an agent’, and they’re experimenting, and that’s fine, but they’re given these tools without a lot of guidance, and the guidance that’s out there is of highly variable quality.”
Eventually, Paterson says, cases will end up in court where legal precedents will be set, and the developers will have a duty to monitor incidents, evolve and improve their protocols in response. ...
See the full story here; https://www.pressreader.com/australia/the-guardian-australia/20260814/282316801863955
OpenAI is AOL
Era-defining economic Bubbles (and bursts) are never caused by random acts of speculation — they are all intentionally inflated. They all start as cultural “phenomena,” invented by “visionaries,” shrouded in “new-age metrics,” and sold to the public by “trusted institutions” as “the next big thing,” until the rug gets pulled and the bottom falls out. ...
The most profound takeaway from these melodramatic macroeconomic stories is that the pops of these Bubbles does not mean the underlying technology is bunk. The internet got pretty big after AOL collapsed, and AI will far outlive the looming AI crash. Rather, the bursts destroy the illusion of the original gateway drugs as an investable business models. As it did for dial up, so it will for AI agents. ...
See the full story here: https://eshap.substack.com/p/openai-is-aol?utm_campaign=email-half-post&r=1vosln&utm_source=substack&utm_medium=email
From Stress-Relief Vibration Patterns to Haptic Emojis, USC Researchers Personalize the Sense of Touch With AI
...
Although personalized haptic feedback could create more intuitive and accessible user experiences, efficiently identifying each person’s preferred vibration patterns has remained a longstanding challenge.
USC haptics researcher Heather Culbertson aimed to close this gap with her latest study.
Led by Culbertson, the research team developed an algorithm called Vibrotactile Preference Learning (VPL), a system that efficiently learns an individual’s preferred vibration patterns for producing a desired emotional response. ...
Individual differences in physiology, including factors such as skin thickness and the density of touch receptors, as well as differences in psychology can significantly influence how people perceive the same haptic signal. This phenomenon is called tactile subjectivity. ...
The researchers created an algorithm called Vibrotactile Preference Learning (VPL), which uses an active querying strategy to identify a user’s preferred vibration pattern in as few as 40 rounds of pairwise comparisons. By comparison, traditional methods may require nearly 200 comparisons to identify a favorite among just 20 vibration signals.
The system first represents each vibration using four key parameters in what researchers call a four-dimensional space: intensity, motor balance (texture), rhythm (pulse frequency) and grain (pulse duty cycle).
VPL then uses a Gaussian process preference model to estimate a user’s preferences across that four-dimensional space. Rather than asking users to compare every possible vibration pair, the algorithm intelligently selects the next pair expected to provide the most useful information. ...
The research could also enable more personalized forms of digital communication through “haptic emojis” — customized vibration patterns that convey emotions such as affection, comfort or excitement. The technology could even create unique haptic signatures for different contacts, allowing users to recognize whether a call is from a partner, parent or close friend without looking at a screen.
See the full story here: https://viterbischool.usc.edu/news/2026/07/from-stress-relief-vibration-patterns-to-haptic-emojis-usc-researchers-personalize-the-sense-of-touch-with-ai/?hash=gmXCHEqg86001
Pages
- About Philip Lelyveld
- Mark and Addie Lelyveld Biographies
- Presentations and articles
- Trustworthy AI – A Market-Driven approach
- Tufts Alumni Bio