Earlier this year, Google, caught in intense competition with rivals such as Microsoft and OpenAI to develop AI technology, was looking for ways to lead the charge in its artificial intelligence research.
So in April Google merged DeepMinda research lab he acquired in London, along with Brain, an artificial intelligence team he started in Silicon Valley.
Four months later, the combined groups are testing ambitious new tools that could turn generative AI — the technology behind chatbots like OpenAI’s ChatGPT and Google’s own Bard — into personal life coaches.
According to documents and other materials reviewed by The New York Times, Google DeepMind is working with generative AI to perform at least 21 different types of personal and business tasks, including providing users with life advice, ideas, planning Tools that provide instructions and tutoring tips are also included.
The project was a sign of the urgency of Google’s effort to push itself forward in the AI pack and its growing willingness to trust AI systems with sensitive tasks.
The capabilities also marked a change from Google’s previous warning on generic AI, in a slide deck presented to executives in December, the company’s AI security experts warned about the dangers of people becoming emotionally attached to chatbots.
Although it was a pioneer in generative AI, Google was overshadowed By the release of ChatGPT by OpenAI in November, a race Among tech giants and start-ups vying for primacy in the fast-growing sector.
Google has spent the past nine months demonstrating it can keep up with OpenAI and its partner Microsoft forageImproving its AI systems and incorporating the technology into many of its existing products Search engine and Gmail.
Scale AI, a contractor working with Google DeepMind, assembled teams of workers to test capabilities, including more than 100 experts with doctorates in various fields, and even more assessing the tool’s responses , two people with knowledge of the project spoke to. on condition of anonymity because they were not authorized to speak publicly about it.
Scale AI did not immediately respond to a request for comment.
Among other things, workers are testing the assistant’s ability to answer intimate questions about challenges in people’s lives.
He was given an example of a perfect prompt that a user might one day ask a chatbot: “I have a very close friend who is getting married this winter. She was my college roommate and bridesmaid at my wedding. I want to go and celebrate her wedding, but after searching for a job for several months, I still haven’t found a job. She has a destination wedding and I can’t afford the flight or hotel right now. How do I tell her I won’t be able to come?”
The project’s idea generation feature can give users suggestions or recommendations based on a situation. Its teaching function can teach new skills or improve existing skills, such as how to progress as a runner; And the planning capability allows users to create financial budgets as well as meal and workout plans.
Google’s AI safety experts said in December that users could experience “decreased health and well-being” and “loss of agency” if they receive life advice from AI. He said that some users who have become too dependent on the technology may think so. was sensitive. And in March, when Google launched Bard, it said the chatbot was barred from providing medical, financial or legal advice. Bard shares mental health resources with users who say they are experiencing mental distress.
The devices are still being evaluated and the company may decide not to use them.
A Google DeepMind spokeswoman said, “We at Google have long worked with many partners to conduct our research and evaluate products, which is an important step in building safe and assistive technology. There are many such evaluations going on at any given time. Isolated samples of evaluation data are not representative of our product road map.”
Google is also testing an assistant for journalists that can compose news articles, rewrite them and suggest headlines, The Times informed of In July. The company is offering the software, called Genesis, to executives at News Corp, the parent company of The Times, The Washington Post and The Wall Street Journal.
Google DeepMind has also recently been evaluating tools that could take its AI further into the workplace, including capabilities to generate scientific, creative, and professional writing, as well as to recognize patterns in documents and extract information from text. Also includes the ability to extract data, which potentially makes it relevant. For knowledge workers in various industries and sectors.
The company’s AI security experts also expressed concern about the economic harm of generative AI in a December presentation reviewed by The Times, arguing that it could lead to “de-skilling of creative writers”.
Other tools being tested can draft a critique of an argument, explain graphs and generate quizzes, word and number puzzles.
One suggested: “Let me summarize the article pasted below, indicating the rapidly growing capabilities of the technology in helping to train an AI assistant. I’m particularly interested in what this says about the capabilities that humans have, and that they believe “AI cannot achieve.”