The safety-critical nature of autonomous vehicle (AV) operation necessitates development of task-relevant algorithms that can reason about safety at the system level and not just at the component level. To reason about the impact of a perception failure on the entire system performance, such task-relevant algorithms must contend with various challenges: complexity of AV stacks, high uncertainty in the operating environments, and the need for real-time performance. To overcome these challenges, in this work, we introduce a Q-network called SPARQ (abbreviation for Safety evaluation for Perception And Recovery Q-network) that evaluates the safety of a plan generated by a planning algorithm, accounting for perception failures that the planning process may have overlooked. This Q-network can be queried during system runtime to assess whether a proposed plan is safe for execution or poses potential safety risks. If a violation is detected, the network can then recommend a corrective plan while accounting for the perceptual failure. We validate our algorithm using the NuPlan-Vegas dataset, demonstrating its ability to handle cases where a perception failure compromises a proposed plan while the corrective plan remains safe. We observe an overall accuracy and recall of 90% while sustaining a frequency of 42Hz on the unseen testing dataset. We compare our performance to a popular reachability-based baseline and analyze some interesting properties of our approach in improving the safety properties of an AV pipeline.

Robots such as autonomous vehicles and assistive manipulators are increasingly operating in dynamic environments and close physical proximity to people. In such scenarios, the robot can leverage a human motion predictor to predict their future states and plan safe and efficient trajectories. However, no model is ever perfect -- when the observed human behavior deviates from the model predictions, the robot might plan unsafe maneuvers. Recent works have explored maintaining a confidence parameter in the human model to overcome this challenge, wherein the predicted human actions are tempered online based on the likelihood of the observed human action under the prediction model. This has opened up a new research challenge, i.e., \textit{how to compute the future human states online as the confidence parameter changes?} In this work, we propose a Hamilton-Jacobi (HJ) reachability-based approach to overcome this challenge. Treating the confidence parameter as a virtual state in the system, we compute a parameter-conditioned forward reachable tube (FRT) that provides the future human states as a function of the confidence parameter. Online, as the confidence parameter changes, we can simply query the corresponding FRT, and use it to update the robot plan. Computing parameter-conditioned FRT corresponds to an (offline) high-dimensional reachability problem, which we solve by leveraging recent advances in data-driven reachability analysis. Overall, our framework enables online maintenance and updates of safety assurances in human-robot interaction scenarios, even when the human prediction model is incorrect. We demonstrate our approach in several safety-critical autonomous driving scenarios, involving a state-of-the-art deep learning-based prediction model.

Recently, an RTI filed by the Internet Freedom Foundation (IFF) revealed that the Delhi Police is using Facial recognition technology (FRT) to nab rioters in the capital city. This has caused an uproar as many members of the civil society raised concerns and called the Delhi Police's use of FRT'unethical' in the absence of a Data Protection Act in the country. The argument being made by them is national security should not come at the cost of privacy. Technology such as FRT has been controversial, and authorities leveraging such tech is definitely a concern. The RTI filed by IFF revealed that the procurement of the FRT by the Delhi Police was authorised as per a 2018 direction of the Delhi High Court in Sadhan Haldar v NCT of Delhi.

Facial-recognition technology (FRT) is mainly deployed in the cybersecurity and surveillance sectors. It has long been in use at airport borders and on smartphones, and as a tool to help police identify criminals. But it is now creeping further into private and public spaces. From Quito to Nairobi, Moscow to Detroit, hundreds of municipalities have installed cameras equipped with FRT, sometimes promising to feed data to central command centres as part of'safe city' or'smart city' solutions to crime. The COVID-19 pandemic might accelerate their spread.

Facebook is almost fully abandoning facial recognition, but its parent company Meta isn't. On November 2, the world's largest social media network said it's going to stop using facial recognition technology (FRT) systems on its platform and delete facial recognition templates for billions of people. However, Meta spokesperson Jason Grosse told Recode that the move doesn't apply to its upcoming metaverse products. The social media firm rebranded to Meta on October 29 when chief executive Mark Zuckerberg announced that the company is shifting its focus to building a future metaverse. "The next platform will be even more immersive -- an embodied internet where you're in the experience, not just looking at it. We call this the metaverse, and it will touch every product we build," Zuckerberg said in a letter following the Facebook Connect event.

Over the last few years, the conversation around emerging technologies like AI and machine learning has increased massively. However, this conversation is limited only to the research and developers' community. The general public, which is at the receiving end, is largely left out of such conversations. This is mainly because there has been very little effort to give cultural and linguistic context to such technologies. To give an example, most of us might be unaware of what AI is called in our local tongue or worse; there might not be any local term to refer to AI to begin with.

The number of US adults teleworking due to the pandemic fell by 30% between January and May 2021 (from 23% to 16%), with the biggest drop in May. As more employees return to their offices, new and unexpected challenges hidden within the new hybrid work model threaten to severely disrupt safety and security. Another added complication is that newly relaxed CDC (Centers for Disease Control) guidelines do not apply to unvaccinated people or account for variations in local or state health guidelines that may be in place, potentially creating a need for organizations to identify and track different groups of employees and visitors. Thankfully, AI-based solutions exist to bolster the security of in-person workplaces and enhance key elements, from the sign-in process to restricted area enforcement, while also allowing for frictionless adherence to health guidelines. As the most accurate, turnkey biometric solution, facial recognition has the potential to solve many of the looming challenges offices will soon face as employees return to work.

The European Commission's (EC) proposed Artificial Intelligence (AI) regulation – a much-awaited piece of legislation – is out. While this text must still go through consultations within the EU before its adoption, the proposal already provides a good sense of how the EU considers the development of AI within the years to come: by following a risk-based approach to regulation. Other use-cases such as FRT for authentication processes are not part of the list of high-level risks and thus should require a lighter level of regulation. While technology providers have to maintain the highest level of performance and accuracy of their systems, this necessary step isn't the most critical to prevent harm. The EC doesn't detail any threshold of accuracy to meet, but rather requires a robust and documented risk-mitigation process designed to prevent harm.

As technology advances, machines are becoming more and more sophisticated, the tools we utilize increasingly intricate, and the currently exceptionally debated and discussed "AI" is being brought into each aspect of our lives. Artificial Intelligence has a vast number of uses in the present society. It is being utilized in aviation, training, computer science, finances, ecology, medical care, heavy industry, marketing, e-commerce business, customer service, and transportation. One could continue for quite a long time this way, only listing the industries in which AI is prominent. Being so widely used to upgrade the professional existences of so many, it is no big surprise that Artificial Intelligence has infiltrated the pursuit of employment and recruitment markets.

Like humans, pigs have idiosyncratic faces, and new players in the Chinese pork market are taking notice, experimenting with increasingly sophisticated versions of facial recognition software for pigs. China is the world's largest exporter of pork, and is set to increase production next year by 9%. As the nation's pork farms grow in scale, more farmers are turning to AI systems like facial recognition technology – known as FRT – to continuously monitor, identify, and even feed their herds. This automated style of farming has the potential to be safer, cheaper and generally more effective: In 2018, pig farmers in China's Guangxi province trialling FRT found that it slashed costs, cut down on breeding time, and improved welfare outcomes for the pigs themselves. But it also has the potential to leave behind independent, small-scale farmers, who cannot afford to introduce this kind of technology to their operations.