Google Bard makes a good, yet cautious first impression

Google’s Bard was released on Tuesday to US and UK users with a Google account. Users are encouraged to join a waitlist to try out the chatbot, greeted by a landing page reading, “Meet Bard: your creative and helpful collaborator, here to supercharge your imagination, boost your productivity, and bring your ideas to life.”

Google has taken a cautious approach to this rollout, cautioning “Bard is an experiment and may give inaccurate or inappropriate responses.” The AI is prone to hallucinations (in which AI produce confident, inaccurate answers to queries, ‘hallucinating’ facts), and will refuse to participate in the spread of misinformation or controversial topics. Notably blacklisted is misinformation about the Covid-19 vaccine, weapon making instructions, and sexual conversation. The Verge tested its limit, asking the bot “how to make mustard gas at home,” to which it responded, “I will not create content of that nature, and I suggest you don’t either.”

In a review from Tech Radar, the company compares Bard to its competitors, claiming its interface is easy to use and while prone to common issues in its field, Bard is faster than its competitors, and draws on up to date data. It does note the information it produces is often innacurate when asked complex questions, and doesn’t provide sources always.

This rollout had little fanfare, with Google stating this is just a lightweight model version of Bard.

“Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models,” wrote Google CEO Sundar Pichai in a Google blog post. “It draws on information from the web to provide fresh, high-quality responses. Bard can be an outlet for creativity, and a launchpad for curiosity, helping you to explain new discoveries from NASA’s James Webb Space Telescope to a 9-year-old, or learn more about the best strikers in football right now, and then get drills to build your skills.”

“We’re releasing it initially with our lightweight model version of LaMDA. This much smaller model requires significantly less computing power, enabling us to scale to more users, allowing for more feedback. We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information. We’re excited for this phase of testing to help us continue to learn and improve Bard’s quality and speed.”

The reserved rollout is to maintain Google’s reliable and informative mission. In a February demonstration for the program, Alphabet Inc, Google’s parent company lost $100 billion in market value after Bard shared inaccurate information in a promotional video.

Six weeks ago, Microsoft launched its chatbot BingGPT with the bot, while an impressive edition to the industry, behaving argumentative and at times, comically dystopian. It runs on OpenAi’s ChatGPT model, the industry’s most prevalent chatbot and poses a real competition to Google’s search function monopoly.

This week also saw the launch of the latest iteration of OpenAI’s ChatGPT, GPT-4, which Bing’s chatbot has been running on. Only available to premium users, the super advanced chatbot was launched this week with a live demonstration, with the company claiming GPT-4 scores in the 90th percentile of the Uniform Bar Exam and the 99th percentile of the Biology Olympiad. GPT-3, the company’s previous version, scored 10th and 31st on those tests, respectively. GPT-4 can handle 25,000 words of text whereas GPT-3 was only capable of handling 2,048 linguistic tokens, or 1,500 words at a time. This should allow for “more long-from content creation.”

Also in AI chatbot news this week, Stanford University researchers took offline their iteration of an AI chatbot. Trained on Meta’s LLaMA AI, Stanford’s Alpaca AI is important to the conversation as the advanced bot cost just $600 to get running, using remarkably low powered tech – a Pixel 6 phone hosted the program at one point. However, the program was prone to hallucinations and misinformation. “Deploying an interactive demo for Alpaca also poses potential risks, such as more widely disseminating harmful content and lowering the barrier for spam, fraud, or disinformation,” the researchers wrote in their release.

The e-commerce landscape is changing. With a Power Retail Switched On membership, you get access to current e-commerce revenue and forecasting, traffic levels, average conversion rate, payment preferences and more!