Google is said to be tinkering with a whole set of new AI tools. These tools should make your life easier and help you in 21 areas of your life.
Planner, teacher and life coach. These are all functions that Google wants generative AI to perform. The company is tinkering with AI tools to power chatbots that can perform 21 tasks.
Safer than Bard?
New research is needed to realize such things with generative AI. Bard and other chatbots currently discourage users to ask such advice. “Do not rely on Bard’s answers as medical, legal, financial or other professional advice,” reads the warning in Bard’s Privacy Notice.
Another tricky element is that such advice can only be given if a user provides personal information. But an AI bot further trains itself on user input. Personal information from other chatbot users can then be retrieved by entering appropriate search queries into the chatbot. This is a huge problem in the face of hackers.
Research into new applications for chatbots would be made possible by the merger of DeepMind and Brain in April. With this, Google wanted to consolidate its AI research divisions and follow up tools from competitors in the market faster.
AI tested out by a jury
The New York Times reports that Google wants to solve the problems by having current generative AI models tested out by a panel of experts. The answers would be analyzed by employees of Scale AI, a data-labelling startup with ties to Google’s DeepMind. In addition, the panel will be joined by more than 100 experts who have expertise in the field of AI.
It is uncertain whether the chatbot will also come to the public or be stored away after the testing phase.