Following Microsoft and Meta into the unknown, AI startup Anthropic - maker of Claude - has a new technique to prevent users from creating or accessing harmful content - aimed at avoiding regulatory
actions against the company, gaining investment funding, and convincing advertisers and other types of businesses it is safe to adopt its models.
The technology identifies whether content is correctly optimized to serve queries and recommendations related to the brand in GAI models such as OpenAI's ChatGPT, Google's Gemini and Meta's Llama.
Shailesh Prakash, vice president and general manager for Google News, has resigned. He is known to be important to the company's relationships with publishers.
Google reportedly paid $2.7 billion to bring back an AI genius as it and competitors continue to build out services built on AI.
Google researchers think they have found a way to reduce hallucination caused by large language models powering today's generative AI technologies.
Improving the learning process that requires less data is a key element of an AI framework developed by Google DeepMind and Imperial College London.
Llama 3.1, created in partnership with Nvidia, is available in three unique versions - including one that marks Meta's biggest open-source AI model yet - keeping Meta in direct competition with tech
giants and startups including Google, Amazon, OpenAI and Anthropic.
Florence-2 - a type of AI called a vision language model - can create content for marketing, advertising, and storytelling by really understanding text prompts to generate images.
Google uses the code and reasoning capabilities of LLMs to replicate human behavior and identify and demonstrate security vulnerabilities. The "Naptime" name refers to software engineers being allowed
to take naps while the software helped to solve problems.
Brave has integrated AI into its search engine to create a real-time answer engine using large language models, data from its search results and retrieval augmented generation to generate answers
almost instantly.
Advertisers try to persuade people to buy products, healthcare providers try to persuade people to make healthier lifestyle changes, and politicians try to persuade people to support their policies
and vote for them. It will become more important as platforms including Google and Microsoft learn how to build persuasiveness into ad models.
Google has unveiled a new family of open-source AI models called Gemma to take on Meta and others, with models capable of running directly on a developer's laptop or desktop computer.
State-affiliated hackers from Russia's military intelligence, Iran's Revolutionary Guard, and Chinese and North Korean governments are using AI tools from OpenAI to improve hacking using large
language models, according to a report published Wednesday.
Two Microsoft researchers set out to determine whether or not developers could make large language models unlearn any part of the data they have been trained on.
Apple has reportedly begun negotiations with major news publishers over permission to use their content as it develops its own generative AI platforms. Contracts would include Conde Nast, NBC News and
IAC, "The New York Times" reported.
Microsoft's new AI model, Phi-2, is capable of running on devices and outperforming others from Meta or Google, the company said Tuesday.
A Google research paper on TW-BERT describes how a new framework improves search rankings without requiring major changes because it integrates with existing query expansion models and improves
performance.
The group's goal is to advance AI safety research, identifying best practices and standards, and facilitate information sharing among policymakers and the industry.
The expanded partnership, announced with the release of the new large language model Llama 2, signals increased collaboration across the industry and an open approach to sharing technology.
The MosaicML research team is expected to join Databricks when the transaction closes, and its technology will be integrated with Databricks' Lakehouse platform.
A group of UC Berkeley students and faculty members conducted an experiment that allowed people to compare two generative AI models simultaneously, giving models the ability to compete for the best
response.
Sebastien Bubeck, lead author on the paper about artificial general intelligence, has documented complex behaviors exhibited by the system during the past several months.
Google has updated math and logic capabilities in its chatbot Bard by incorporating advancements developed in its Pathways Language Model, a large language model similar to GPT.
UBS analysts examined the cost risks for Alphabet around the integration of generative AI into Google search and have seen recent evidence that costs are coming down rapidly, including OpenAI stating
that it had managed to reduce costs for ChatGPT by 90% over December through March.
Google's Universal Speech Model is a critical first step toward creating AI that can understand and translate 1,000 languages. Earlier this week, Google shared details of its AI universal speech
model designed to understand hundreds of spoken languages.
HubSpot is launching two OpenAI-powered tools today to help marketing, sales and customer service professionals "work smarter, not harder" -- Content Assistant, to help rapidly create and share
quality content, and ChatSpot.ai to complete tasks using a natural-language chat-based user experience.
Meta released its own AI language generator LLaMA on Friday. Meta claims LLaMA is "desirable" in an industry led by large language models because of the minimal computing power it requires to test
new approaches, further research in the field and explore new use cases for AI.
One new tool is a multilingual email collector that lets users to send email or SMS invitations in local languages.
Meta's AI unit built a model the company says is the first to translate 200 different languages with state-of-the-art quality validated through extensive evaluations for each.
Kids believe their generation has a much clearer understanding of gender compared to older generations, with 80% of kids and teens trusting people their age to understand who they are and their
gender, including how to articulate and express it. It has begun to surface in search results. The outcome will create challenges if not faced now.