Google’s announcement of Bard comes amid widespread excitement about OpenAI’s ChatGPT, which was released less than three months ago. Last month, reports surfaced that the launch of ChatGPT raised concerns among Google executives, as the chatbot threatens the company’s dominant position in the internet search space. Bard appears to be Google’s answer to ChatGPT. The service, powered by Google’s LaMDA conversational language model, will offer more insightful responses to questions compared to the current “quick” answers that Google Search provides.
AI-Powered Features on Google Search
Unlike ChatGPT — at least in its current state — Google’s Bard will be connected to the internet. It will be able to draw on multiple perspectives when there’s no single answer and break down complex information into a simple, easily understandable format, Pichai said. “Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models. It draws on information from the web to provide fresh, high-quality responses,” he explained. Google is set to add new AI-Powered features to Google Search. The company continues to invest heavily in AI technologies, which are “creating entirely new ways to engage with information, from language and image to video and audio,” Pichai said. “We’re working to bring the latest AI advancements into our products, starting with Search.” Last month, The New York Times revealed that Google is planning to release 20 new AI-powered products this year. Google’s immediate goal is to get feedback on the service and ensure its responses meet quality and safety requirements. Therefore, it is releasing Bard on a “lightweight” model version of LaMDA for the time being. “This much smaller model requires significantly less computing power, enabling us to scale to more users, allowing for more feedback,” Pichai said. Pichai said Google would start inviting developers to experiment with its Generative Language API in March. The company plans to create a suite of APIs and tools to enable developers to create innovative AI-based apps easily.
‘Developing AI Responsibly’
Pichai reaffirmed Google’s commitment to developing AI in a responsible manner and according to its AI principles published in 2018. These principles include making AI systems that do not create or reinforce unfair biases in society and limiting the potential that they can be adapted for potentially harmful purposes. Experts have warned that generative AI systems could maintain biases in the materials used to train them and in the information they gather online. In December, Steven Piantadosi from the University of California, Berekly’s Computation and Language Lab, shared a series of tweets showing ChatGPT reinforcing discriminatory biases. In one instance, Piantadosi asked ChatGPT to write a Python program to determine whether a person should be tortured based on their country of origin. The system wrote a code that specifically selected people from North Korea, Syria, and Iran for torture. Cybercriminals are also taking advantage of ChatGPT to further their nefarious schemes. There have been reports of cybercriminals using ChatGPT to write code for malware and create phishing emails. Last month, Check Point researchers revealed that cybercriminals are using ChatGPT to create hacking tools and new dark web markets. It is unclear how Google plans to tackle these and other shortcomings of ChatGPT.
