SXSW this year covered the electric car and automated vehicles in the streets. Waymo, [Google]Alphabet’s automated car, has been providing some rideshare services and mapping in the streets of downtown Austin for a little while. This shiny governance & policy object attracted the attentions of Illinois governor JB Pritzker and South Carolina Governor Henry McMaster.
That meant in some cases bomb sweep technicians and secret State service skirted quickly around some of the SXSW AI sessions. The more controversial sessions covered (AI enabled) mental health applications using brain chips (BCIs) to manage people who are depressed, expedite US pill regulatory conventions using a combination of calculations on molecules and the poor medical ethics of Chinese human testing (most likely using Falun Gong and other Chinese dissidents). Journalist complained about the deletion of key library and training search terms* in AI automation used by DEI dictat. [*They can always reassert word terms in a different context, because language without government subsidy is legal in 1st Amendment law.]While other developers tried to manage consumer use case questions about how data is training models for education purposes, cyber investigations and surveillance.
However, the biggest auditorium magnets for AI was fair use and intellectual property policy voiced at the crisis inflection point to preserve human culture. John Hegel, author of The Journey Beyond Fear, gave a standing room only talk about opportunity models for data intermediaries to hedge against IP theft and abuse of copyrighted material. AI legal speakers for Andreesen & Horowitz, public affairs and VC capital for AI small innovators [Y combinator], indicated a decentralized pathway out of Tech feudal oligarchies in emerging AI policy. Small technologists as innovators are dependent now on support of antitrust, in a lowered regulatory environment.
SEE ALSO:
* TURN YOUR DATA INTO ASSETS (Responsibly) at Personal Digital Spaces
* Thomson Reuters scores early win in AI copyright battles in the US
* Thomson Reuters v. Ross Intelligence: Copyright, Fair Use, and AI (Round One)
* Court: Training AI Model Based on Copyrighted Data Is Not Fair Use as a Matter of Law
Small businesses are the self employment alternative to being a worker facing HR’s integration of AI tools. Most of AI-for-work developers shared how the tools work from LLMs and AI Agents are meant to shed highly detailed but mercilessly repetitive tasks like: data entry, repeat calculations analysis after a digital render, find & replace functions and more. While some project an anticipation to the end of all work by AI replacement, others see paths forward to unencumbered freedom from tasks that demand too much time for low yield of results. Others credibly fear being discarded and micromanaged under a cloud of transactional estimation from bots trained on biased material priorities by their bosses. The trick for labor disputes is to negotiate the Black Box boundary where training data layer transparency can be traced back to demands for specific programming prompts by the business against notorious operations ‘tenting’ or siloed project management. This would suss out bids to perform functions later found to be quantifiably harmful, transgressive of personal rights or simply biased against human interest.
SEE ALSO: OpenAI Says DeepSeek Could ‘Be Compelled’ to Cause Harm, Calls for Ban
You began to hear a common thread as technologists and creative artists demanded means from one another to provide cloud housing for technical security and IP legal enforcement to stave off AI poaching of now and future works. Complications with Europe splitting off AI innovation as current strategy, given DSA and AI Act regulations, were discussed with some bracing and some of a wait-&-see approach. Antitrust is seen as a given reliable path to ensuring no one national government can command all of the IP capital or creative equity and personal data at the same time.
SEE ALSO: EU Lawmakers Push Back on U.S. Criticism of Tech Antitrust Regulation
One thing is for sure, Tesla has an American electronic auto competitor in Rivian. Their 100% American made electric car sales start at $80,000. The big sale point for us is Rivian’s ability to manage privacy settings in electronic dashboard to switch off GPS and data sharing inside the cars. They’re now competition, not only to Tesla, but GMC and other American car makers who install heavy surveillance capability inside newer models with no way to shut involuntary data sharing down. It was Obama-era electric car makers who chose to sell private locating [GIS] information to global government “marketplace exchanges” to insure margins on their products, not the US consumer. That’s how a 2012 dumb Subaru became a coveted car, not necessarily for safety or performance. For those who want an electric car, you can have privacy & American made certainty in a vehicle that works.
THE MUSIC OF SXSW ‘25
This conference is most known for producing the best music showcases. Here’s the sample select cuts/ best witness of SXSW ‘25.
Texas & Americana
Electronic & Synth Pop
Liberty in Many Directions is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit sheilamdean.substack.com/subscribe