Google DeepMind is developing more than 20 new generative AI features

According to reports, Google is developing a series of generative artificial intelligence features that can perform tasks such as providing life advice and setting budgets

The New York Times quoted Google’s internal According to the filing, the initiative is reportedly being led by Google DeepMind, the artificial intelligence research group of search giant Google. The team was formed in April this year after Google merged two early artificial intelligence teams, DeepMind and Google Brain.

It is reported that Google engineers are working on at least 21 different generative AI functions, and one of the main highlights is the ability to provide suggestions for "challenges in people's lives." Information tips obtained by The New York Times indicate that the feature can handle complex user queries containing multiple lines of text.

According to reports, in December last year, Google’s artificial intelligence safety experts allegedly issued a warning that machine-generated life suggestions could have a negative impact on users. The Times reports that it's uncertain whether Google will roll out the feature, along with other generative AI features detailed in the report. Those AI features are expected to focus on a variety of use cases. One of the features in development is described as an AI-based tutor, while another is designed to help users create planning documents such as personal budgets. It is said that as part of the plan, Google DeepMind will also develop "creative creation capabilities."

Google is reportedly working on the development with well-funded AI startup Scale AI. Scale AI sells tools that companies can use to create training datasets and integrate large language models into their software. Scale AI's valuation reached US$7.3 billion after completing a round of financing in early 2021

According to reports, Scale AI has established multiple teams to assist Google in testing its new artificial intelligence capabilities. In addition to other employees, these teams reportedly include more than 100 experts with Ph.D.

Google DeepMind said in a statement: "We have long-term cooperation with multiple partners to evaluate our work at Google. research and products, which is a critical step in building safe, useful technology. We have been conducting multiple evaluations, and isolated data samples are not representative of our product planning."

Google is also actively advancing many other generation Plans for artificial intelligence

According to the latest reports, Google is developing a large language model combination called Gemini. It is reported that Gemini can generate text and images and is planned to be launched this fall. Gemini is expected to be available to users in the form of a new chatbot or as an upgraded version of Google's existing Bard AI assistant

The above is the detailed content of Google DeepMind is developing more than 20 new generative AI features. For more information, please follow other related articles on the PHP Chinese website!