The HBA panel on AI included, from left, Perla Villarreal, Peter N. Salib, Aly Dossa, Dwight Silverman and George Earle. Photo courtesy of the HBA.
June 02, 2023 – University of Houston Law Center professors joined experts in the field of technology and media to discuss the intersection of big tech and free speech as part of the Houston Bar Association President’s Speaker Series. The half-day conference was held at John M. O’Quinn Law Building at the University of Houston Law Center at the end of May.
The first panel on “Artificial Intelligence & Free Speech: The Usefulness and Challenges of AI” was moderated by Perla Villarreal, Technology Consultant at Thoughtworks and Houston Chapter Lead for Latinas in Tech.
Artificial Intelligence & Alignment
Panelists included Dwight Silverman, Tech Columnist for the Houston Chronicle, Peter N. Salib, Assistant Professor at the University of Houston Law Center, George Earle, Global Commercial Director at Thoughtworks and Aly Dossa, Shareholder and Chair of the Intellectual Property Practice and the Data Security & Privacy Practice at Chamberlain Hrdlicka.
The panelists delved into the challenges posed by AI and possible impacts on free speech, which is protected by the First Amendment.
University of Houston Law Center Assistant Professors, from top, Peter N. Salib Nikolas Guggenberger and Aman Gebru commenting during panel discussions about big tech and free speech.
The fight currently before the courts is whether content used to train an AI model is fair use or commercial, said Dossa who expects to see a “new licensing scheme that will compensate the original source generators.”
Panelists also spoke on the importance of establishing guardrails and regulations to guide AI applications.
On how to control AI content, Earle said that competition would not be enough. “It’s going to have to be regulations, which means law.”
Salib added that in the AI regulatory space, we are not ready and “don’t know what regulations we need.”
Silverman noted one of the guardrails in place now is limiting a user’s time with AI.
“What you can hope is that as the AI models get better, more capable, that the companies creating them get more nuanced and better at their guardrails,” said Silverman.
“The fundamental nature of deep learning, on which large language models are based, is that we do not know why they do the things they do, and so we do not know if the guardrail works until it breaks,” said Salib. “If you are worried about a no guardrails AI, especially as AI gets more capable, then you should be worried about AI with guardrails because the problem of controlling AI is a problem we have not solved.”
The panel discussed the rapid pace at which AI is evolving, drawing comparisons to the early days of the internet and social media.
“I have never seen such an important technology become so quickly available to the masses as this. It took years and years before most people could afford a PC,” said Silverman, who has been writing about technology since the mid-1980s. “Everyone has been using ChatGTP since November.”
“There is gold rush going on,” said Earle. “It’s like the internet was in the early 90s in Silicon Valley.”
On whether AI will ultimately be beneficial or detrimental, Salib noted the tech industry is working on what is called the Alignment Problem, trying to align AI with human values and reasoning.
“The bad news is that compared with the number of dollars invested in making AI smarter and more powerful” the money invested in aligning AI is “several orders of magnitude lower,” said Salib.
Social Media & Section 230
University of Houston Law Center Assistant Professors Nikolas Guggenberger and Aman Gebru along with Lisa Falkenberg, Vice President and Editor of Opinion at the Houston Chronicle, spoke on the second panel about “Social Media: How Private Ownership, Federal Mandates, and Court Rulings Intersect with the First Amendment.”
The social media panel, moderated by KPRC Consumer Investigative Reporter Amy Davis, also touched on regulations and cited Section 230 from the 1996 Communications Decency Act, which, according to the DOJ, “provides immunity to online platforms from civil liability based on third-party content and for the removal of content in certain circumstances.”
“With respect to section 230, I think reform is overdue,” said Guggenberger about the nearly 30-year-old statue. “Section 230 was created for two reasons. One is to encourage moderation, and two, to protect small players as they compete with large platforms. Both of these are obsolete.”
Gebru recommended “tweak the edges to improve [Section 230]” rather than doing a complete overhaul. We can learn from the intellectual property system. “The cutting-edge technology when IP laws were set up was the printing press.” I think we “keep the basic rules intact, but experiment with different systems before adopting major changes.”
The HBA conference included 9 panel speakers and a keynote by Eric Schnapper, Professor of Law at the University of Washington and attorney for the plaintiffs in the U.S. Supreme Court case Gonzalez v. Google. The case challenged the protections given to online platforms under Section 230. The Supreme Court, in a per curiam decision in May, remanded the case to the Ninth Circuit without addressing the question of Section 230.
The event was cohosted by the Houston Bar Association Law & the Media Committee and the Houston Pro Chapter of the Society of Professional Journalists.
For more on the HBA President’s Speakers Series, click here.
Amy Davis, KPRC Reporter, left, moderated the panel on social media at the HBA event with speakers Lisa Falkenberg, Houston Chronicle, Aman Gebru, Assistant Professor University of Houston Law Center and Nikolas Guggenberger, Assistant Professor University of Houston Law Center. Photo courtesy of the HBA.
