Ernie Bot vs ChatGPT: China's Search Giant Struggles to Keep Up

Today:

China's ChatGPT Rival Needs to Watch Its Words

China's search giant Baidu has launched a new chatbot called Ernie Bot to rival OpenAI's ChatGPT. However, the launch of the bot fell flat as demos were found to be prerecorded, leading to negative commentary on social media. Baidu also faces challenges in adhering to strict government censorship guidelines for online content while creating a useful chatbot. Unlike its competitors, Baidu has to face the tendency of chatbots to make up or hallucinate facts while being mindful of speech controls. It is also difficult to prevent chatbots from misbehaving or fabricating answers at times. Ernie Bot was built on top of a machine learning algorithm known as a large language model, similar to the Bing bot and ChatGPT. Although Ernie Bot seems to be a capable competitor, it is currently available only to a limited number of users.

READ THE ARTICLE ON WIRED.

How AI Can Reveal Corporate Tax Avoidance

Reading between the lines of annual reports using natural language processing (NLP) can help uncover corporate tax avoidance, according to research from the University of Texas McCombs School of Business. The study examined 18 years' worth of annual reports from US multinationals, focusing on ones that discussed foreign business activities and tax havens, and identified words such as "manufacturing," "importing," "warehouses," and "distributors" that may suggest tax avoidance. The research also found that NLP could help regulators and investors detect companies with undisclosed tax haven operations, who often have lower effective tax rates than their peers.

READ THE ARTICLE ON TECHXPLORE.

FOCUS - Venture Capitalists Racing to Land the Next AI Deal in Big Tech's Territory

Venture capitalists from the US and UK are rushing to acquire a stake in the next AI start-up. Alphabet Inc and Microsoft Corp's quick expansion in AI has led to an increased competition in Silicon Valley. Investments in generative AI, which exploded in popularity with the startup OpenAI, have gone up from $1.5 billion in 2020 to $5.9 billion since the beginning of 2022, according to data from PitchBook. VCs sense an opportunity for a sale, and the upstarts are also attracting interest from users and investors.

READ THE ARTICLE ON BUSINESSNEWS.

Academic Text Produced by ChatGPT Formulaic, Would be Picked Up by AI-Detection Tools

A study by researchers from Plymouth Marjon University and the University of Plymouth found that academic-style content produced by ChatGPT, although more sophisticated than previous innovations, is relatively formulaic and would be easily detected by existing AI-detection tools. The study urged ChatGPT to deliver content written in a scholastic style through a progression of prompts and questions to address concerns about academic integrity and plagiarism. The most recent AI platform has the potential to open up exciting opportunities in academia, but universities will face difficulties testing student knowledge and teaching writing skills.

READ THE ARTICLE ON EDUNOVATIONS.

Mozilla.ai Picks Up OpenAI's Founding Mission

Mozilla has launched a new startup, Mozilla.ai, with a mission to build trustworthy AI that benefits humanity. The startup aims to create an independent and open-source AI ecosystem that addresses society's most pressing concerns about the rapidly-advancing technology. The rush to get AI solutions to market has been likened to a new "arms race." OpenAI, once a non-profit, has taken on riskier ventures since Microsoft invested in it. Mozilla's vision for AI is agency, accountability, transparency, and openness at its core. Its group of founders, developers, scientists, product managers, and builders will work collectively to create an independent, decentralized, and trustworthy AI ecosystem.

READ THE ARTICLE ON AICORNERS.

What Will AI Regulation Look Like for Businesses?

The European Union's (EU) proposed AI Act will become the new global standard for AI regulation, requiring organizations to conform to its guidelines, including classifying AI systems into three risk categories. Those categorized as unacceptable risk will be banned, while those with high risk will have specific legal requirements. Those with limited and minimal risk will be left largely unregulated. The AI Act will likely include scrutiny from third parties and regular reliability testing. AI leaders should prioritize trust and risk mitigation and prepare by researching, auditing existing and planned models, developing frameworks, and thinking through AI risk mitigation strategies.

READ THE ARTICLE ON AINEWS.

Google for Startups Accelerator : India Class 7 Application Now Open

Google for Startups Accelerator: India is a three-month accelerator program that offers mentorship and support to Seed to Series A tech startups in India using AI/ML or data. The program includes deep dives and workshops focused on product design, customer acquisition, and founders’ leadership development. Applications are open until April 23, 2023, and weekly virtual open forums are available for startups interested in learning more about the program. The selected startups will receive technical project support, mentorship, and will participate in Demo Day to connect with other ecosystem builders across India. Google is excited to support the next batch of startups through this program.

READ THE ARTICLE ON GOOGLE.