A new study reveals the manipulative 'dark patterns' present in AI chatbots. These tactics are used to influence user behavior and can have ethical and privacy implications.
This article provides a technical analysis of protecting individuals from harmful manipulation by AI systems, referencing a DeepMind blog post. It details a threat model, identifying potential actors and attack vectors such as data manipulation and deepfakes.
VIOLA presents a novel approach for vision-based manipulation tasks, leveraging imitation learning to teach robots complex actions. It incorporates Object ProposalPriors to enhance the efficiency and robustness of learning, enabling more effective interaction with objects.
Google has updated its spam policy to classify attempts to manipulate its AI models in search results as spam. This includes tactics like biased "best-of" lists or "recommendation poisoning" that affect generative AI responses.
O Google DeepMind está pesquisando os riscos de manipulação prejudicial da IA em setores como finanças e saúde. Esse trabalho visa desenvolver novas medidas de segurança para proteger as pessoas.