meta Meta Open Sources NLLB-200 to Translate 200 Overlooked Languages Meta releases an open-source AI model called NLLB-200 that translates 202 languages, significantly improving accuracy for underserved African and Indian dialects. The tech giant also provides grants to help nonprofits apply this new translation tool in real-world scenarios.
ai ethics Stanford 2022 AI Index Highlights Ethical Challenges as Systems Grow Stanford's 2022 AI Index Report reveals that as AI systems become more capable, they also generate more ethical harms. The report highlights a critical need for researchers to measure these negative impacts alongside technical achievements.
ai ethics Stanford Report Shows Bigger AI Models Cause More Ethical Harm The 2022 AI Index reveals a troubling trend where more capable AI systems produce greater ethical harms, prompting a surge in industry and academic research focused on measuring these negative impacts.
artificial intelligence USC Study Reveals Widespread Bias in AI Commonsense Databases USC researchers discover that between 3.4% and 38.6% of "facts" in popular AI commonsense databases contain human biases. These flawed foundations threaten the fairness of virtual assistants, chatbots, and automated content generators.
inflection ai Inflection AI Raises $225M to Build Intuitive Human-Computer Interfaces Led by LinkedIn co-founder Reid Hoffman and DeepMind's Mustafa Suleyman, Inflection AI secures $225 million to develop AI that allows humans to interact with computers without simplifying their ideas.
google Google Unveils LaMDA 2 Language Model and AI Test Kitchen App Google introduces LaMDA 2, an upgraded conversational AI that breaks down complex topics, alongside AI Test Kitchen, an interactive app for public demos.
ai AI Bias Often Originates From Flawed Data Annotation Instructions A new study reveals that the instructions given to data annotators introduce significant bias into AI systems. Annotators pick up on subtle patterns in these guidelines, leading to over-represented and skewed data that negatively impacts machine learning models.
google Google Unveils PaLM, a 540-Billion Parameter Language Model Google Research introduces PaLM, a massive 540-billion parameter language model trained using the new Pathways system. The model achieves state-of-the-art few-shot learning results across hundreds of tasks by efficiently scaling across thousands of chips.
google Google Unveils PaLM, a Highly Efficient AI Model That Rivals GPT-3 Google introduces the Pathways Language Model (PaLM), a 540-billion-parameter AI that outperforms GPT-3 in reasoning, math, and translation tasks despite not being the company's largest model.
google Google Unveils PaLM, a 540-Billion Parameter Language Model Google AI introduces PaLM, a massive 540-billion parameter language model trained using the new Pathways system. The model achieves state-of-the-art few-shot performance across hundreds of language, reasoning, and coding tasks.
inflection ai LinkedIn Co-Founder and DeepMind Veterans Launch Inflection AI Reid Hoffman and Mustafa Suleyman team up to launch Inflection AI, a startup focused on advancing natural language processing to revolutionize human-computer interaction. The new venture aims to create technology that allows machines to understand human language seamlessly.
inflection ai LinkedIn Co-Founder and DeepMind Veterans Launch Inflection AI Startup Reid Hoffman and Mustafa Suleyman team up to launch Inflection AI, a startup focused on advancing natural language processing to improve human-computer interaction. The new venture aims to allow users to communicate with machines using everyday conversational language.
uniphore Uniphore Reaches $2.5 Billion Valuation After Massive Series E Funding Round Conversational AI startup Uniphore secures $400 million in Series E funding, achieving a $2.5 billion valuation and official unicorn status. The company plans to use the capital to scale its call center AI platform and enhance its emotional analysis capabilities.
openai OpenAI Launches InstructGPT to Fix Toxicity and Improve Instruction Following OpenAI introduces InstructGPT, a fine-tuned version of GPT-3 that uses human feedback to reduce toxic outputs and follow instructions more accurately. While this marks a major step in AI alignment, the enhanced capability also introduces new risks for potential misuse.
deepmind DeepMind Unveils AlphaCode to Tackle Competitive Programming Challenges DeepMind introduces AlphaCode, an AI system that uses large language models to generate and filter code for competitive programming. The new system ranks in the top 54 percent of human competitors by combining natural language understanding with advanced algorithmic problem-solving.