The main goal of this paper is to study how often cookie banners that comply with the General Data Protection Regulation (GDPR) contain aesthetic manipulation, a design tactic to draw users' attention to the button that permits personal data sharing. As a byproduct of this goal, we also evaluate how frequently the banners comply with GDPR and the recommendations of national data protection authorities regarding banner designs. We visited 2,579 websites and identified the type of cookie banner implemented. Although 45% of the relevant websites have fully compliant banners, we found aesthetic manipulation on 38% of the compliant banners. Unlike prior studies of aesthetic manipulation, we use a computer vision model for salient object detection to measure how salient (i.e., attention-drawing) each banner element is. This enables the discovery of new types of aesthetic manipulation (e.g., button placement), and leads us to conclude that aesthetic manipulation is more common than previously reported (38% vs 27% of banners). To study the effects of user and/or website location on cookie banner design, we include websites within the European Union (EU), where privacy regulation enforcement is more stringent, and websites outside the EU. We visited websites from IP addresses in the EU and from IP addresses in the United States (US). We find that 13.9% of EU websites change their banner design when the user is from the US, and EU websites are roughly 48.3% more likely to use aesthetic manipulation than non-EU websites, highlighting their innovative responses to privacy regulation.

LLM web agents now browse and take actions on the open web, yet current agent evaluations are constrained to sandboxed environments or artificial tasks. We introduce BrowserArena, a live open-web agent evaluation platform that collects user-submitted tasks, runs Arena-style head-to-head comparisons, and uses step-level human feedback to surface failure modes. Collecting and analyzing step-level annotations on the agent traces, we identify three consistent failure modes: captcha resolution, pop-up banner removal, and direct navigation to URLs. By constructing targeted datasets to further study these tasks, we discover variations in how different language models navigate these failure modes. We find, for example, that o4-mini deploys a wider variety of strategies to circumvent captcha resolution than other models and DeepSeek-R1 consistently misleads users about pop-up banner closure. Our findings surface both the diversity and brittleness of current web agents. More broadly, our benchmarking methodology provides an approach to evaluating and understanding web agent failure modes at scale.

Crafting a marketing message (copy), or copywriting is a challenging generation task, as the copy must adhere to various constraints. Copy creation is inherently iterative for humans, starting with an initial draft followed by successive refinements. However, manual copy creation is time-consuming and expensive, resulting in only a few copies for each use case. This limitation restricts our ability to personalize content to customers. Contrary to the manual approach, LLMs can generate copies quickly, but the generated content does not consistently meet all the constraints on the first attempt (similar to humans). While recent studies have shown promise in improving constrained generation through iterative refinement, they have primarily addressed tasks with only a few simple constraints. Consequently, the effectiveness of iterative refinement for tasks such as copy generation, which involves many intricate constraints, remains unclear. To address this gap, we propose an LLM-based end-to-end framework for scalable copy generation using iterative refinement. To the best of our knowledge, this is the first study to address multiple challenging constraints simultaneously in copy generation. Examples of these constraints include length, topics, keywords, preferred lexical ordering, and tone of voice. We demonstrate the performance of our framework by creating copies for e-commerce banners for three different use cases of varying complexity. Our results show that iterative refinement increases the copy success rate by $16.25-35.91$% across use cases. Furthermore, the copies generated using our approach outperformed manually created content in multiple pilot studies using a multi-armed bandit framework. The winning copy improved the click-through rate by $38.5-45.21$%.

We introduce LLMmap, a first-generation fingerprinting technique targeted at LLM-integrated applications. LLMmap employs an active fingerprinting approach, sending carefully crafted queries to the application and analyzing the responses to identify the specific LLM version in use. Our query selection is informed by domain expertise on how LLMs generate uniquely identifiable responses to thematically varied prompts. With as few as 8 interactions, LLMmap can accurately identify 42 different LLM versions with over 95% accuracy. More importantly, LLMmap is designed to be robust across different application layers, allowing it to identify LLM versions--whether open-source or proprietary--from various vendors, operating under various unknown system prompts, stochastic sampling hyperparameters, and even complex generation frameworks such as RAG or Chain-of-Thought. We discuss potential mitigations and demonstrate that, against resourceful adversaries, effective countermeasures may be challenging or even unrealizable.

Online advertising banners are sold in real-time through auctions.Typically, the more banners a user is shown, the smaller the marginalvalue of the next banner for this user is. This fact can be detected bybasic ML models, that can be used to predict how previously won auctionsdecrease the current opportunity value. However, learning is not enough toproduce a bid that correctly accounts for how winning the current auctionimpacts the future values. Indeed, a policy that uses this prediction tomaximize the expected payoff of the current auction could be dubbedimpatient because such policy does not fully account for the repeatednature of the auctions. Under this perspective, it seems that most biddersin the literature are impatient. Unsurprisingly, impatience induces a cost.We provide two empirical arguments for the importance of this cost ofimpatience. First, an offline counterfactual analysis and, second, a notablebusiness metrics improvement by mitigating the cost of impatience withpolicy learning

Text-to-image models such as stable diffusion have opened a plethora of opportunities for generating art. Recent literature has surveyed the use of text-to-image models for enhancing the work of many creative artists. Many e-commerce platforms employ a manual process to generate the banners, which is time-consuming and has limitations of scalability. In this work, we demonstrate the use of text-to-image models for generating personalized web banners with dynamic content for online shoppers based on their interactions. The novelty in this approach lies in converting users' interaction data to meaningful prompts without human intervention. To this end, we utilize a large language model (LLM) to systematically extract a tuple of attributes from item meta-information. The attributes are then passed to a text-to-image model via prompt engineering to generate images for the banner. Our results show that the proposed approach can create high-quality personalized banners for users.

A judge has found "reasonable evidence" that Elon Musk and other executives at Tesla knew that the company's self-driving technology was defective but still allowed the cars to be driven in an unsafe manner anyway, according to a recent ruling issued in Florida. Palm Beach county circuit court judge Reid Scott said he'd found evidence that Tesla "engaged in a marketing strategy that painted the products as autonomous" and that Musk's public statements about the technology "had a significant effect on the belief about the capabilities of the products". The ruling, reported by Reuters on Wednesday, clears the way for a lawsuit over a fatal crash in 2019 north of Miami involving a Tesla Model 3. The vehicle crashed into an 18-wheeler truck that had turned on to the road into the path of driver Stephen Banner, shearing off the Tesla's roof and killing Banner. The lawsuit, brought by Banner's wife, accuses the company of intentional misconduct and gross negligence, which could expose Tesla to punitive damages. The ruling comes after Tesla won two product liability lawsuits in California earlier this year focused on alleged defects in its Autopilot system.

The sun had yet to rise in Delray Beach, Fla., when Jeremy Banner flicked on Autopilot. His red Tesla Model 3 sped down the highway at nearly 70 mph, his hands no longer detected on the wheel. Seconds later, the Tesla plowed into a semi-truck, shearing off its roof as it slid under the truck's trailer. Banner was killed on impact. Banner's family sued after the gruesome 2019 collision, one of at least 10 active lawsuits involving Tesla's Autopilot, several of which are expected to go to court over the next year. Together, the cases could determine whether the driver is solely responsible when things go wrong in a vehicle guided by Autopilot -- or whether the software should also bear some of the blame.

We've all taken automatic language translation for granted, especially where web pages are concerned. But Google has finally added it to where it might be the most useful: the mobile version of Gmail. To be fair, automatic translation has been a feature of Gmail for years, but only on the web. Now, Google is building it into its mobile app, where millions of people access their email every day. Android users will see automatic translation within their Gmail app via an update scheduled beginning today, August 8; iOS users will receive an update as early as August 21.

Assuming Chrome does not push back its removal of cookies, 2023 will likely be the end of the already constrained cookie-based targeting era. Growing solutions such as universal IDs and first-party publisher data show promise, but the reality is that the adtech industry will also need to refine the ability to target and measure performance in the total absence of user-level signals – a task AI can help with. One of the more promising solutions for this future lies in AI's ability to create lookalike models for brands based on smaller sets of known users that do perform well. In some ways, AI has already been creating lookalike modeling for cookie-based targeting, just using a different set of audience data that will soon be much more scarce such as demographics, behaviors and interests. The good news is there are many privacy-compliant signals available including the type of website or app, geography, type of device, time of day, local weather, dominant political or other attributes of the region, keywords on the page, sentiment of the page and time on page.