Last month, I found myself in a rather extraordinary position. Seated just a few spots away from me was none other than Elon Musk. On the other side of the table, I had the pleasure of sharing the space with Bill Gates. And scattered throughout the room were other prominent figures such as Satya Nadella, the CEO of Microsoft, and Eric Schmidt, the former CEO of Google. Situated at the opposite end of the table was Sam Altman, the head of OpenAI, the company responsible for ChatGPT. We had all gathered for the inaugural meeting of Senate Leader Chuck Schumer’s AI Insight Forum, a series of events aimed at expediting the development of bipartisan legislation on artificial intelligence.
The attendees encompassed a diverse group, consisting of senators, tech executives, civil-society representatives, and myself—an UC Berkeley computer-science researcher entrusted with the task of bringing the wealth of academic knowledge on AI accountability to the forefront.
Looking back, I can’t help but question what was truly accomplished in that room. Much of the discussion seemed to revolve around hypothetical concerns and possibilities, rather than addressing the present reality. It’s perfectly acceptable to speculate about the future, as long as we don’t become consumed by daydreams. Unfortunately, it seems that American lawmakers are entering the realm of tangible AI regulation while neglecting to gain a clear understanding of the current landscape.
One of the challenges in these conversations lies in the broad usage of the term “AI” itself. It has become one of those marketing buzzwords that can be twisted and molded to fit various narratives. According to Congress’ own classification, AI encompasses everything from basic risk assessments to facial-recognition tools, from automated decision-making systems to deepfake political images, from recommendation algorithms on online platforms to seemingly intelligent chatbots. The term “AI” simply refers to a model that follows a predetermined path from input to output, bypassing the need for human intervention and relying on computer calculations instead.
As with any popular buzzword in the business world, “AI” is heavily leveraged in technology advertising. At the forum, executives extolled its extraordinary capabilities. AI was presented as a force that could revolutionize education, cure diseases like cancer, eradicate poverty and hunger, supercharge productivity, and transform the workforce. Of course, these proclamations were coupled with warnings of grave dangers, with some attendees even expressing concerns about AI being weaponized by malicious actors or leading to global disasters if it fell into the wrong hands. Elon Musk described AI as a “double-edged sword,” an incredibly powerful technology that could bring about immediate catastrophe if misused.
As the meeting was closed to the press, the exact details of what took place that day remain undisclosed. Naturally, everyone was eager to know what transpired during the closed-door discussions. They were particularly interested in the insights shared by Musk and Altman. Following the meeting, some senators criticized the lack of transparency, while Schumer echoed the views put forward by the tech executives and praised the meeting’s success.
Undeniably, AI is powerful and its potential dangers are real. However, as these perspectives echo through committee hearings, government advisory boards, press releases, and lobbying efforts, it is apparent that focusing solely on influential corporate voices is a limited approach. It’s tempting to construct artificial contexts or extrapolate possibilities rather than observe the reality that AI is already impacting everyone’s lives. I speak from experience in academic circles, where discussions often revolve around theoretical social and legal theories or complex mathematical equations and code repositories. Many researchers, whether using words or symbols, tend to speak in general terms and hypothetical scenarios. Data sets are frequently divorced from context or meaning, and proper documentation remains a chronic issue. The benchmarks we rely on to evaluate AI models often have no connection to real-world applications and consequences.
Ensuring the safety of millions of Americans necessitates a more grounded perspective. During Schumer’s forum, Laura MacCleery from the Latino-advocacy group UnidosUS shared her experience with previous tech initiatives in education, recounting the story of a broken computer monitor being used as a doorstop in a low-income school district. Similar anecdotes from civil-rights organizations and labor-union leaders reminded me of the multifaceted nature of the situation. While AI holds the potential to alleviate poverty, it also leaves people vulnerable to financial scams. It may advance cancer research, but struggles to produce meaningful outcomes in healthcare. AI can boost productivity in the workplace, but it also brings about precarious job positions such as AI raters and rampant piracy.
We’ve witnessed instances where AI systems didn’t perform as expected in real-world settings. In recent years, I’ve read accounts of AI systems revealing their limitations and fallibility, rather than being the mythical, sentient beings portrayed in popular culture. For example, a pregnant Black woman, Porcha Woodruff, was wrongfully arrested due to a false facial-recognition match. Brian Russell had to fight for years to clear his name after being falsely accused of unemployment fraud by an algorithm. Tammy Dobbs, an elderly woman with cerebral palsy, lost vital home care due to algorithmic mishaps. And Davone Jackson found himself locked out of the low-income housing he desperately needed to escape homelessness due to a false flag triggered by an automated tenant-screening tool.
“They didn’t choose this,” Fabian Rogers, a tenant organizer in Brooklyn, once said to me. The residents in his public-housing building were in a dispute with their landlord over the use of facial recognition in the new security system. “The hardest part is explaining to someone who is struggling to pay rent and put food on the table why they should care about any of this,” he added.
I have come to realize what Rogers meant. The inaugural forum led by Schumer was not a platform for serious policy deliberation. No corporate secrets were divulged, and the day consisted mostly of softball questions and prepared statements. Throughout my years of advocacy and research, I have often found myself in similar advisory panels, surrounded by decision-makers in comfortable office chairs, while gazing out the window at an enticing stretch of lush greenery beyond the beige curtains. As always, we spent the day shifting slightly in our cushioned swivel chairs.
The truth is, “AI” itself is an elusive concept. While the technology is very much real, the term itself is intangible. It can either be the enthusiastic pitch of a marketing executive or the weary sigh of someone grappling with the consequences of minute engineering decisions that have disrupted their entire life. As lawmakers finally begin to take action on AI, we all have a choice about whose voices we listen to.
Denial of responsibility! Vigour Times is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
