This article was co-published with Understanding AI, a newsletter that explores how A.I. works and how it's changing our world. When Silicon Valley executives testify before Congress, they normally get raked over the coals. But OpenAI CEO Sam Altman's Tuesday appearance before the Senate Judiciary Committee went differently. Senators asked Altman probing questions and listened respectfully to his answers. Afterward, the committee's chairman, Sen. Richard Blumenthal of Connecticut praised Altman.

At a Senate hearing on Tuesday, the CEO of OpenAI Sam Altman received a warm welcome from lawmakers, many of whom expressed surprise at his main argument: that AI should be regulated, and fast. It was a far cry from the grueling ordeals that tech CEOs have previously faced on Capitol Hill. Mark Zuckerberg, Jack Dorsey and Shou Zi Chew have all endured antagonistic Senate hearings in recent years about the wide-ranging impacts of their platforms--Facebook, Twitter and TikTok, respectively--on American democracy and the lives of their users. "I think what's happening today in this hearing room is historic," said Senator Dick Durbin (D., Ill.) during the Senate judiciary subcommittee hearing about oversight of AI. "I can't recall when we've had people representing large corporations or private sector entities come before us and plead with us to regulate them." But in calling for legal guardrails to govern the tech his company is building, Altman is not unlike the other Silicon Valley leaders who have testified before Congress in the past.

For weeks after his bizarre conversation with Bing's new chatbot went viral, New York Times columnist Kevin Roose wasn't sure what had happened. "The explanations you get for how these language models work, they're not that satisfying," Roose said at one point. "No one can tell me why this chatbot tried to break up my marriage." He's not alone in feeling confused. Powered by a relatively new form of AI called large language models, this new generation of chatbots defies our intuitions about how to interact with computers.

'Media Buzz' host Howard Kurtz join'Sunday Night in America with Trey Gowdy' to discuss a poll claiming Americans are blaming the media for divisiveness in America. Be afraid, be very afraid. That's the message that is starting to dominate the media's many channels when it comes to artificial intelligence. And it's not just prognosticators but such voices as Elon Musk and the grandfather of AI that are saying an apocalyptic future may loom in the distance. I'm not hitting the panic button yet, but the sheer velocity of what AI is either able to achieve or is moving toward achieving seems to increase exponentially each week.

OpenAI CEO Sam Altman, the artificial intelligence lab behind ChatGPT, took questions from reporters after his congressional hearing, including his definition of "scary AI." OpenAI CEO Sam Altman testified before Congress in Washington, D.C., this week about regulating artificial intelligence as well as his personal fears over the tech and what "scary" AI systems means to him. Fox News Digital asked OpenAI's wildly popular chatbot, ChatGPT, to also weigh in on examples of "scary" artificial intelligence systems, and it reported six hypothetical instances of how AI could become weaponized or have potentially harmful impacts on society. When asked by Fox News Digital on Tuesday after his testimony before a Senate Judiciary subcommittee, Altman gave examples of "scary AI" that included systems that could design "novel biological pathogens." "An AI that could hack into computer systems," he continued. "I think these are all scary. These systems can become quite powerful, which is why I was happy to be here today and why I think this is so important."

Sen. Marsha Blackburn, R-Tenn., shares her takeaways from Tuesday's AI hearing with OpenAI CEO Sam Altman. She also reveals what next steps she and her colleagues are prepared to take to protect consumer data amid the AI boom. EXCLUSIVE: Nashville musicians are increasingly worried about complications with artificial intelligence's growing sophistication that could threaten their livelihood, Sen. Marsha Blackburn, R-Tenn., warned this week. "We met with the Nashville Technology Council a couple of weeks ago, and we have talked with so many of the musicians. They're concerned that using AI, they will do a copycat of their voice and take the lyrics of their song, which you can get on ChatGPT," Blackburn told Fox News Digital during an interview in her Senate office.

In recent years, understanding the decision-making process of black-box models has become not only a legal requirement but also an additional way to assess their performance. However, the state of the art post-hoc interpretation approaches rely on synthetic data generation. This introduces uncertainty and can hurt the reliability of the interpretations. Furthermore, they tend to produce explanations that apply to only very few data points. This makes the explanations brittle and limited in scope. Finally, they provide scores that have no direct verifiable meaning. In this paper, we present BELLA, a deterministic model-agnostic post-hoc approach for explaining the individual predictions of regression black-box models. BELLA provides explanations in the form of a linear model trained in the feature space. Thus, its coefficients can be used directly to compute the predicted value from the feature values. Furthermore, BELLA maximizes the size of the neighborhood to which the linear model applies, so that the explanations are accurate, simple, general, and robust. BELLA can produce both factual and counterfactual explanations. Our user study confirms the importance of the desiderata we optimize, and our experiments show that BELLA outperforms the state-of-the-art approaches on these desiderata.

With the development of large language models (LLMs), zero-shot learning has attracted much attention for various NLP tasks. Different from prior works that generate training data with billion-scale natural language generation (NLG) models, we propose a retrieval-enhanced framework to create training data from a general-domain unlabeled corpus. To realize this, we first conduct contrastive pretraining to learn an unsupervised dense retriever for extracting the most relevant documents using class-descriptive verbalizers. We then further propose two simple strategies, namely Verbalizer Augmentation with Demonstrations and Self-consistency Guided Filtering to improve the topic coverage of the dataset while removing noisy examples. Experiments on nine datasets demonstrate that REGEN achieves 4.3% gain over the strongest baselines and saves around 70% of the time compared to baselines using large NLG models. Besides, REGEN can be naturally integrated with recently proposed large language models to boost performance.

AI-based software systems are rapidly becoming essential in many organizations [1]. However, the focus on the technical side of building artificial intelligence (AI)-based systems are most common, and many projects, more often than not, fail to address critical human aspects during the development phases [2, 3]. These include but are not limited to age, gender, ethnicity, socio-economic status, education, language, culture, emotions, personality, and many others [4]. Ignoring human-centered aspects in AI-based software tends to produce biased and non-inclusive outcomes [5]. Shneiderman [6] emphasizes the dangers of autonomy-first design in AI and the hidden biases that follow. Misrepresenting human aspects in requirements for model selection and data used in training AI algorithms can lead to discriminatory decision procedures even if the underlying computational processes were unbiased [7]. For example, a study by Carnegie Mellon revealed that women were far less likely to receive high-paying job ads from Google than men [8] due to the under-representation of people of color and women in high paying IT jobs. Studies on human-centered design aim to develop systems that put human needs and values at the center of software development and clearly understand the context of the software system's usage [2, 9].

Setting proper evaluation objectives for explainable artificial intelligence (XAI) is vital for making XAI algorithms follow human communication norms, support human reasoning processes, and fulfill human needs for AI explanations. In this position paper, we examine the most pervasive human-grounded concept in XAI evaluation, explanation plausibility. Plausibility measures how reasonable the machine explanation is compared to the human explanation. Plausibility has been conventionally formulated as an important evaluation objective for AI explainability tasks. We argue against this idea, and show how optimizing and evaluating XAI for plausibility is sometimes harmful, and always ineffective in achieving model understandability, transparency, and trustworthiness. Specifically, evaluating XAI algorithms for plausibility regularizes the machine explanation to express exactly the same content as human explanation, which deviates from the fundamental motivation for humans to explain: expressing similar or alternative reasoning trajectories while conforming to understandable forms or language. Optimizing XAI for plausibility regardless of the model decision correctness also jeopardizes model trustworthiness, because doing so breaks an important assumption in human-human explanation that plausible explanations typically imply correct decisions, and vice versa; and violating this assumption eventually leads to either undertrust or overtrust of AI models. Instead of being the end goal in XAI evaluation, plausibility can serve as an intermediate computational proxy for the human process of interpreting explanations to optimize the utility of XAI.