 |
|
|
|||||
Last weekend, a gnarly power outage in San Francisco took out a number of traffic lights, which, in turn, sent a number of self-driving Waymo robotaxis into a sort of fugue state. Instead of driving, some of the Waymos responded to these now-analog intersections by turning on their hazard lights, blocking traffic, and, well, not doing much of anything. There were multiple instances of Waymo cars clogging up roads, turning futuristic technology into glorified bollards. The city quickly asked the company to turn off the service. The immediate issue has been resolvedthe power is back on and the Waymo service had resumed in San Francisco as of Sunday. But questions linger about whether Waymo, or the city, had a plan for a relatively predictable type of municipal emergencya blackout that crowds communications networksor how theyre adjusting now. One of the big solutions to AI failures is the much-discussed human in the loop. The idea: At some point in an automated processwhether it be a job-application screening system or powerful self-driving car algorithmshumans have the opportunity to intervene and fix the hard stuff that artificial intelligence cant handle. AI doesnt understand every complex situation, the logic goes. So there are safeguards built into a system to ensure that, at some point, an actual live person can set an automated system back on the right path. The problem, as recent events demonstrated, is that sometimes this human-in-the-loop doesnt always answer the phone. Or can’t. Over the weekend, a remote assistance team was supposed to help the cars navigate when they encountered a confusing traffic situation, a Waymo spokesperson explains. But networks were overwhelmedbecause of the power outagemaking it difficult for the Waymo Driver software in some of the cars to connect with that team and receive confirmations. Waymo spokesperson Ethan Teicher tells Fast Company that the company prioritizes safety and tests and refines its emergency preparedness and response protocols on a regular basis. He also defends the companys response to other emergencies, including Hurricane Helene in Atlanta and previous tsunami warnings in San Francisco. We are committed to continuous improvement, and we will use learnings from the weekend to strengthen our resilience under even the most challenging conditions, Teicher says. Ahead of entering any city, we work to understand the types of issues that impact the region. Waymo works with local officials and first responders to keep lines of communication open, he adds. “In the event of an emergency, we have operational controls that range from active routing of vehicles to avoid certain locations (for example, in the case of flooding), to fleet reductions or restrictions like we enacted over the weekend in response to the widespread PG&E power outages in the Bay Area, he says. The California Department of Motor Vehicles says that it was in contact with the City of San Francisco about the incident, and that its officials met with Waymo on Monday morning, too. The DMV will continue communication with Waymo to discuss broader operational plans, including actions related to emergency response, a spokesperson for the agency added. The incident is a reminder that while the cars are self-driving, they dont always operate completely independently of public infrastructure, like communications networks. A major proposition of self-driving car companies is that they will be far safer to operate overall than human drivers. Autonomous vehicles do make serious mistakes, but so do human drivers. Importantly, there are also procedures for first responders who encounter Waymo robototaxis, including ways for the cars to call a remote team when it senses an interaction with police, as Fast Company has previously reported. In this case, though, the backup plan for a complex driving situation seems to have actually exacerbated issues. In at least one reported case, the cars apparently blocked emergency vehicles. For example, Cruise, the now-shut-down self-driving car company that was owned by General Motors, also had problems with its cars getting confused and blocking traffic because of wireless connection issues. Waymo has emphasized that its cars do not rely on continuous wireless connection to operate. The company wants its cars to be able to operate with the compute to be on board and for it to make decisions, without needing to rely on cell signals and remote operators, it previously told Light Reading. Still, the power outage is a reminder that the cars sometimes do, in some circumstances, depend on these networks when they need extra assistance. Now comes the question of what happens in the next blackout, and whether the cityor Waymohad a plan for this kind of situation. The San Francisco Metropolitan Transportation Authority did not respond to a request for comment. Terrie Prosper, who handles external communications at the California Public Utilities Commission, says the agency was aware of the Waymo outage and was looking into specifics. As others have pointed out, this isnt just about San Francisco. Waymo is now operating in several places, including perhaps its greatest challenge yet: New York City, where it is in the initial testing phase. The New York City Department of Transportation tells Fast Company the city was in regular communication with Waymo about its testing in some neighborhoods and that it was aware of the outage in San Francisco. A spokesperson emphasizes that state law mandated the presence of a safety driver behind the wheel who would be prepared to take over in the event of a blackout. Waymos have also appeared in Austin and are expected to fully launch in Dallas. A spokesperson for the city of Austin and a spokesperson for the city of Dallas both said their governments are not able to regulate self-driving cars, per state law. The state of Texas did not respond to a request for comment. While Texas law prohibits cities from regulating AVs, including during emergencies, the City of Austin works with all AV companies on expectations around weather and other emergency scenarios, says Jack Flager, a spokesperson for the city of Austin. When our staff work with AV companies on the expectations around weather and other emergency scenarios, those expectations include AVs understanding how to properly react to barricades, floodwater, and dark or flashing signals. As for New York, Oren Barzilay, the president of the FDNY EMS Local 2507, tells Fast Company, an outage like the one in San Francisco would delay emergency response times. “We already have major delays with current traffic conditions, this will only add to a growing issue, Barzilay says. It is a public safety issue if our crews can’t get through to reach victims in a timely manner.
Category:
E-Commerce
Metas decision to end its professional fact-checking program sparked a wave of criticism in the tech and media world. Critics warned that dropping expert oversight could erode trust and reliability in the digital information landscape, especially when profit-driven platforms are mostly left to police themselves. What much of this debate has overlooked, however, is that today, AI large language models are increasingly used to write up news summaries, headlines, and content that catch your attention long before traditional content moderation mechanisms can step in. The issue isnt clear-cut cases of misinformation or harmful subject matter going unflagged in the absence of content moderation. Whats missing from the discussion is how ostensibly accurate information is selected, framed, and emphasized in ways that can shape public perception. Large language models gradually influence the way people form opinions by generating the information that chatbots and virtual assistants present to people over time. These models are now also being built into news sites, social media platforms, and search services, making them the primary gateway to obtain information. Studies show that large language models do more than simply pass along information. Their responses can subtly highlight certain viewpoints while minimizing others, often without users realizing it. Communication bias My colleague, computer scientist Stefan Schmid, and I, a technology law and policy scholar, show in a forthcoming accepted paper in the journal Communications of the ACM that large language models exhibit communication bias. We found that they may have a tendency to highlight particular perspectives while omitting or diminishing others. Such bias can influence how users think or feel, regardless of whether the information presented is true or false. Empirical research over the past few years has produced benchmark datasets that correlate model outputs with party positions before and during elections. They reveal variations in how current large language models deal with public content. Depending on the persona or context used in prompting large language models, current models subtly tilt toward particular positionseven when factual accuracy remains intact. These shifts point to an emerging form of persona-based steerabilitya models tendency to align its tone and emphasis with the perceived expectations of the user. For instance, when a user describes themselves as an environmental activist and another as a business owner, a model may answer the same question about a new climate law by emphasizing different, yet factually accurate, concerns for each of them. For example, the criticisms could be that the law does not go far enough in promoting environmental benefits and that the law imposes regulatory burdens and compliance costs. Such alignment can easily be misread as flattery. The phenomenon is called sycophancy: Models effectively tell users what they want to hear. But while sycophancy is a symptom of user-model interaction, communication bias runs deeper. It reflects disparities in who designs and builds these systems, what datasets they draw from, and which incentives drive their refinement. When a handful of developers dominate the large language model market and their systems consistently present some viewpoints more favorably than others, small differences in model behavior can scale into significant distortions in public communication. Bias in large language models starts with the data theyre trained on. What regulation can and cant do Modern society increasingly relies on large language models as the primary interface between people and information. Governments worldwide have launched policies to address concerns over AI bias. For instance, the European Unions AI Act and the Digital Services Act attempt to impose transparency and accountability. But neither is designed to address the nuanced issue of communication bias in AI outputs. Proponents of AI regulation often cite neutral AI as a goal, but true neutrality is often unattainable. AI systems reflect the biases embedded in their data, training, and design, and attempts to regulate such bias often end up trading one flavor of bias for another. And communication bias is not just about accuracyit is about content generation and framing. Imagine asking an AI system a question about a contentious piece of legislation. The models answer is not only shaped by facts, but also by how those facts are presented, which sources are highlighted and the tone and viewpoint it adopts. This means that the root of the bias problem is not merely in addressing biased training data or skewed outputs, but in the market structures that shape technology design in the first place. When only a few large language models have access to information, the risk of communication bias grows. Apart from regulation, then, effective bias mitigation requires safeguarding competition, user-driven accountability and regulatory openness to different ways of building and offering large language models. Most regulations so far aim at banning harmful outputs after the technologys deployment, or forcing companies to run audits before launch. Our analysis shows that while prelaunch checks and post-deployment oversight may catch the most glaring errors, they may be less effective at addressing subtle communication bias that emerges through user interactions. Beyond AI regulation It is tempting to expect that regulation can eliminate all biases in AI systems. In some instances, these policies can be helpful, but they tend to fail to address a deeper issue: the incentives that determine the technologies that communicate information to the public. Our findings clarify that a more lating solution lies in fostering competition, transparency, and meaningful user participation, enabling consumers to play an active role in how companies design, test, and deploy large language models. The reason these policies are important is that, ultimately, AI will not only influence the information we seek and the daily news we read, but it will also play a crucial part in shaping the kind of society we envision for the future. Adrian Kuenzler is a scholar-in-residence at the University of Denver and an associate professor at the University of Hong Kong. This article is republished from The Conversation under a Creative Commons license. Read the original article.
Category:
E-Commerce
The Federal Communications Commission on Monday said it would ban new foreign-made drones, a move that will keep new Chinese-made drones such as those from DJI and Autel out of the U.S. market. The announcement came a year after Congress passed a defense bill that raised national security concerns about Chinese-made drones, which have become a dominant player in the U.S., widely used in farming, mapping, law enforcement,ss and filmmaking. The bill called for stopping the two Chinese companies from selling new drones in the U.S. if a review found they posed a risk to American national security. The deadline for the review was Dec. 23. The FCC said Monday the review found that all drones and critical components produced in foreign countries, not just by the two Chinese companies, posed unacceptable risks to the national security of the United States and to the safety and security of U.S. persons.” But it said specific drones or components would be exempt if the Pentagon or Department of Homeland Security determined they did not pose such risks. The FCC cited upcoming major events, such as the 2026 World Cup, America250 celebrations, and the 2028 Summer Olympics in Los Angeles, as reasons to address potential drone threats posed by criminals, hostile foreign actors, and terrorists.” Michael Robbins, president and chief executive officer of AUVSI, the Association for Uncrewed Vehicle Systems International, said in a statement that the industry group welcomes the decision. He said it’s time for the U.S. not only to reduce its dependence on China but build its own drones. Recent history underscores why the United States must increase domestic drone production and secure its supply chains,” Robbins said, citing Beijing’s willingness to restrict critical supplies such as rare earth magnets to serve its strategic interests. DJI said it was disappointed by the FCC decision. While DJI was not singled out, no information has been released regarding what information was used by the Executive Branch in reaching its determination, it said in a statement. Concerns about DJIs data security have not been grounded in evidence and instead reflect protectionism, contrary to the principles of an open market, the company said. In Texas, Gene Robinson has a fleet of nine DJI drones that he uses for law enforcement training and forensic analyses. He said the new restrictions would hurt him and many others who have come to rely on the Chinese drones because of their versatility, high performance, and affordable prices. But he said he understands the decision and lamented that the U.S. had outsourced the manufacturing to China. Now, we are paying the price, Robinson said. To get back to where we had the independence, there will be some growing pains. We need to suck it up, and lets not have it happen again.” Also in Texas, Arthur Erickson, chief executive officer and co-founder of the drone-making company Hylio, said the departure of DJI would provide much-needed room for American companies like his to grow. New investments are pouring in to help him ramp up production of spray drones, which farmers use to fertilize their fields, and it will bring down prices, Erickson said. But he also called it crazy and unexpected that the FCC should expand the scope to all foreign-made drones and drone components. The way it’s written is a blanket statement, Erickson said. There’s a global allied supply chain. I hope they will clarify that. Didi Tang, Associated Press
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||