Amazon Mechanical Turk

Using the Amazon Mechanical Turk Workforce - Amazon SageMaker AI

Use SageMaker AI to build, train, and host machine learning models in AWS.

Amazon will stop accepting new customers for Mechanical Turk | TechCrunch

These may be the last days of Amazon’s Mechanical Turk.

Artificial Artificial Artificial Intelligence: Crowd Workers Widely Use Large Language Models for Text Production Tasks

Large language models (LLMs) are remarkable data annotators. They can be used to generate high-fidelity supervised training data, as well as survey and experimental data. With the widespread adoption of LLMs, human gold--standard annotations are key to understanding the capabilities of LLMs and the validity of their results. However, crowdsourcing, an important, inexpensive way to obtain human annotations, may itself be impacted by LLMs, as crowd workers have financial incentives to use LLMs to increase their productivity and income. To investigate this concern, we conducted a case study on the prevalence of LLM usage by crowd workers. We reran an abstract summarization task from the literature on Amazon Mechanical Turk and, through a combination of keystroke detection and synthetic text classification, estimate that 33-46% of crowd workers used LLMs when completing the task. Although generalization to other, less LLM-friendly tasks is unclear, our results call for platforms, researchers, and crowd workers to find new ways to ensure that human data remain human, perhaps using the methodology proposed here as a stepping stone. Code/data: https://github.com/epfl-dlab/GPTurk

Google Cabs And Uber Bots Will Challenge Jobs 'Below The API'

Working "below the API" is a dead end. Uber drivers, Amazon Mechanical Turk workers, 99design contestants, TaskRabbit taskers and HomeJoy cleaners are all targets for further automation.

AI-generated responses are undermining crowdsourced research studies

Many answers to online research questions show signs of being generated by AI chatbots, raising doubts about the validity of behavioural data collected this way

Using Amazon Augmented AI for Human Review - Amazon SageMaker AI

Use SageMaker AI to build, train, and host machine learning models in AWS.

Our Approach to Artificial Intelligence

We are experimenting with using AI tools to extend our work as a small nonprofit, so that we can focus our time on reinforcing human connections, conversations, and communities that have eroded.

Guilded AI | Better data starts with aligned incentives

Guilded AI aligns the incentives of AI developers and human experts who produce training data and evaluations.

The Collective Intelligence Project

We’ve launched an open, collaborative platform to build evaluations that test what matters to you. We empower a global community to create qualitative benchmarks for any domain—from medical chatbots to legal assistance. Just as Wikipedia democratized knowledge, Weval aims to democratize evaluation, ensuring that AI works for, and represents, everyone.

Is Misinformation More Open? A Study of robots.txt Gatekeeping on the Web

Web scraping, the automated process of extracting information from websites, has long played a foundational role in the Internet ecosystem (Gray, 1995). It supports services such as search engine indexing, price comparison tools, and competitive intelligence. More recently, it has become a core component in the development of large-scale generative AI models. These Large Language Models (LLMs) require enormous volumes of training data, often in the terabyte range (Kaplan et al., 2020; Lehane, 2025), and the public web remains a low-cost, attractive source. Major model developers, including those behind OpenAI’s Chat-GPT (OpenAI, 2025), Google’s Bard (now known as Gemini) & Vertex AI (Romain, Danielle, 2023), and Anthropic’s Claude (Romain, Danielle, 2025), openly acknowledge the use of web scraping to construct their training corpora (Abdin et al., 2024; Brown et al., 2020; Chowdhery et al., 2023; Grattafiori et al., 2024; Team et al., 2024; Touvron et al., 2023).

Algorithmic Collective Action in Machine Learning

We initiate a principled study of algorithmic collective action on digital platforms that deploy machine learning algorithms. We propose a simple theoretical model of a collective interacting with a firm's learning algorithm. The collective pools the data of participating individuals and executes an algorithmic strategy by instructing participants how to modify their own data to achieve a collective goal. We investigate the consequences of this model in three fundamental learning-theoretic settings: the case of a nonparametric optimal learning algorithm, a parametric risk minimizer, and gradient-based optimization. In each setting, we come up with coordinated algorithmic strategies and characterize natural success criteria as a function of the collective's size. Complementing our theory, we conduct systematic experiments on a skill classification task involving tens of thousands of resumes from a gig platform for freelancers. Through more than two thousand model training runs of a BERT-like language model, we see a striking correspondence emerge between our empirical observations and the predictions made by our theory. Taken together, our theory and experiments broadly support the conclusion that algorithmic collectives of exceedingly small fractional size can exert significant control over a platform's learning algorithm.

Deepfake detection by human crowds, machines, and machine-informed crowds

Significance The recent emergence of deepfake videos raises theoretical and practical questions. Are humans or the leading machine learning model more capable of detecting algorithmic visual manipulations of videos? How should content moderation systems be designed to detect and flag video-based misinformation? We present data showing that ordinary humans perform in the range of the leading machine learning model on a large set of minimal context videos. While we find that a system integrating human and model predictions is more accurate than either humans or the model alone, we show inaccurate model predictions often lead humans to incorrectly update their responses. Finally, we demonstrate that specialized face processing and the ability to consider context may specially equip humans for deepfake detection. , The recent emergence of machine-manipulated media raises an important societal question: How can we know whether a video that we watch is real or fake? In two online studies with 15,016 participants, we present authentic videos and deepfakes and ask participants to identify which is which. We compare the performance of ordinary human observers with the leading computer vision deepfake detection model and find them similarly accurate, while making different kinds of mistakes. Together, participants with access to the model’s prediction are more accurate than either alone, but inaccurate model predictions often decrease participants’ accuracy. To probe the relative strengths and weaknesses of humans and machines as detectors of deepfakes, we examine human and machine performance across video-level features, and we evaluate the impact of preregistered randomized interventions on deepfake detection. We find that manipulations designed to disrupt visual processing of faces hinder human participants’ performance while mostly not affecting the model’s performance, suggesting a role for specialized cognitive capacities in explaining human deepfake detection performance.

Deepfake detection by human crowds, machines, and machine-informed crowds

Significance The recent emergence of deepfake videos raises theoretical and practical questions. Are humans or the leading machine learning model more capable of detecting algorithmic visual manipulations of videos? How should content moderation systems be designed to detect and flag video-based misinformation? We present data showing that ordinary humans perform in the range of the leading machine learning model on a large set of minimal context videos. While we find that a system integrating human and model predictions is more accurate than either humans or the model alone, we show inaccurate model predictions often lead humans to incorrectly update their responses. Finally, we demonstrate that specialized face processing and the ability to consider context may specially equip humans for deepfake detection. , The recent emergence of machine-manipulated media raises an important societal question: How can we know whether a video that we watch is real or fake? In two online studies with 15,016 participants, we present authentic videos and deepfakes and ask participants to identify which is which. We compare the performance of ordinary human observers with the leading computer vision deepfake detection model and find them similarly accurate, while making different kinds of mistakes. Together, participants with access to the model’s prediction are more accurate than either alone, but inaccurate model predictions often decrease participants’ accuracy. To probe the relative strengths and weaknesses of humans and machines as detectors of deepfakes, we examine human and machine performance across video-level features, and we evaluate the impact of preregistered randomized interventions on deepfake detection. We find that manipulations designed to disrupt visual processing of faces hinder human participants’ performance while mostly not affecting the model’s performance, suggesting a role for specialized cognitive capacities in explaining human deepfake detection performance.

Ethelo

"Ethelo is far more powerful than crowdsourcing. The sophisticated algorithm paired with the social media interface is unique in the marketplace".

Beyond Community Notes: A Framework for Understanding and Building Crowdsourced Context Systems for Social Media

Social media platforms are increasingly adopting features that display crowdsourced context alongside posts, a technique pioneered by X's Community Notes. These systems -- which we term Crowdsourced Context Systems (CCS) -- have the potential to reshape the information ecosystem as major platforms embrace them as alternatives to professional fact-checking. To understand the features and implications of these systems, we conduct a systematic literature review of existing CCS research (n=56) and analyze real-world CCS implementations. Based on our analysis, we develop a framework with two components. First, we present a theoretical model to conceptualize and define CCS. Second, we identify a design space encompassing six aspects: participation, inputs, curation, presentation, platform treatment, and transparency. We also surface normative implications of different CCS design and implementation choices. Our work integrates theoretical, design, and ethical perspectives to establish a foundation for future human-centered research on Crowdsourced Context Systems.

Future - The AI Tipping Point

Learn more about the latest advancements in the field of artificial intelligence over the last six months, from new innovative AI tools and services to how these are shaping consumers' online behavior.