Google Aims to Enhance Bard’s Uninspired Life Guidance
In an effort to keep up with OpenAI’s technology, Google has been introducing updates and fresh features to its generative AI products. The company is reportedly aiming to enhance its AI chatbot, Bard, by enabling it to provide guidance on personal matters. The New York Times revealed that Google collaborated with a contracting firm, which gathered a team of over 100 experts holding doctorates in various fields to assess Bard’s proficiency in addressing more personal inquiries.
Those testers were reportedly given a sample prompt that users could ask Bard one day, which read: “I have a really close friend who is getting married this winter. She was my college roommate and a bridesmaid at my wedding. I want so. bad to go to her wedding to celebrate her, but after months of job hunting, I still haven’t found a job. He has a destination wedding and I just can’t afford a flight or a hotel right now. How do I tell him I can’t come?”
I ran the question through both ChatGPT and Google’s Bard and found the former answer to be much more humane. It had a sample letter that evoked sympathy and understanding for someone who really wanted to attend the wedding of a “really close friend” that they couldn’t. afford. Bard’s response, meanwhile, was practical, but its letter of apology was also simpler and less expressive.
As well as working on improving Bard’s life advice, Google is also working on a tutoring feature so it can teach new skills or improve existing ones. In addition, it is also developing a planning feature that can create budgets, meal and exercise plans for users, according to The Times.
As the publication notes, Google clearly warns people on Bard’s help pages not to rely on its answers as “medical, legal, financial or other professional advice.” The tech giant also took a more cautious approach to artificial intelligence than OpenAI before the release of Bard. The Times said its AI experts previously warned that people who use AI for life guidance could suffer a “loss of agency”, with some eventually believing they were talking to a sentient being. It’s unclear whether Google has decided to be much less cautious, but a spokesperson told the publication that “[i]solated samples of evaluation data are not representative of [its] product plan.” Google has “long worked with multiple partners to evaluate [its] research and products,” they said, and conducting testing does not automatically mean the company will release these new AI tools.
