Google's DeepMind is using generative artificial intelligence to build a life advice robot.

2023-08-18

Would you accept computer advice? Yes, we all search for information on computers, but Google's DeepMind department is developing something more extensive. According to reports, the company is developing 21 different generative AI tools for providing life advice and coaching. This may signify Google's highest level of trust in generative AI systems. Documents obtained by The New York Times reveal the efforts made by DeepMind to turn AI chatbots into advice providers. In a test prompt, the chatbot was asked how someone could tell a friend that they couldn't attend a wedding due to financial issues. This is a heavy situation that requires an AI capable of generating illusions and fictional facts. However, DeepMind is reportedly trying to steer the project away from "therapeutic" purposes and make it more general - like asking advice from a newspaper advice columnist, except this columnist is a robot. The project's creative generation module can provide users with advice or recommendations based on given situations, and the coaching feature can offer plans for learning new skills. The AI planning component can create meal plans, family budgets, and more. Google has partnered with ScaleAI, a startup aimed at testing and validating AI systems. It is reported that over 100 staff members with various doctoral degrees are testing the advice bot. Information reviewed by The New York Times shows that Google's AI experts have expressed serious concerns about the use of chatbots. In a demonstration in December last year, security experts claimed that consumers might inappropriately rely on chatbots, leading to a "decline in health and happiness levels." However, Google seems to be increasingly abandoning its previous cautious approach to generative AI. Google's AI researchers have invented transformation models that support chatbots like ChatGPT and Google Bard. The company has used this technology internally, but generative AI has not evolved into a consumer product due to accuracy and abuse issues. OpenAI initially expressed concerns about the consequences as well but eventually partnered with Microsoft to provide ChatGPT. The integration of AI with Bing caught Google off guard, resulting in the hasty launch of Bard. Since the AI boom earlier this year, we have seen these chatbots provide misleading and potentially dangerous information, create new malware, and unintentionally infringe on copyrighted content. Nevertheless, DeepMind still believes that generative AI can be guided to provide useful advice.

RECENT AI NEWS

RECENT AI TOOLS