AI will be a central part of Google's I/O event on Wednesday when the tech giant is expected to reveal its PaLM 2 large language model and AI tools for its Workspace apps. A theme of the conference is how AI is "helping people reach their full potential." According to various reports, Google could announce more generative AI experiences and updates for Search and Bard, its AI chatbot. More: - According to CNBC, Google has been working on updates for Bard, including a "Multi-Bard" version that can code and answer math problems.
- The report says Google is also testing "Big Bard" and "Giant Bard" without providing further details.
- At I/O, Google may also announce new AI features related to its search engine, including a possible AI chatbot under a project code-named Magi.
Workspace apps: - Google has already said it would roll out generative AI tools to most or all of its Workspace products, which include Gmail, Chat, Meet, Calendar, Drive, Docs, Sheets, Slides, Forms, and Sites.
- On Wednesday, Google is expected to demo how AI can generate templates in Google Sheets and images in Slides and Meet.
- It began testing AI-enabled writing features in Gmail and Google Docs earlier this year.
- On Friday, Google made its Bard AI chatbot available to users with Google Workspace accounts.
Zoom out: - Google's AI tools are built upon its large language model known as PaLM (Pathways Language Model).
- According to CNBC, Google will announce PaLM 2, which includes 100+ languages, on Wednesday.
- In addition to writing, the LLM, known internally as "Unified Language Model," can also do coding and math, the report states.
- Last month, Google launched Med-PaLM 2, a large language model that answers medical questions for the healthcare sector, to select cloud customers for testing.
CNBC 
| |
What happened: India's Ministry of Electronics and IT has decided not to pursue any AI regulations at this time, a move that could offer a boost to the country's shortage of AI-first startups. Details: India's government views AI as a game-changer for entrepreneurs and businesses in the country. Unlike the European Union or China, India has decided to take a hands-off approach to AI, calling it a "significant and strategic" area for the South Asian market. What the numbers say: According to a recent study from KPMG and The University of Queensland, people in India are the most likely to see AI regulations as unnecessary. While most respondents in 17 other countries said they believe AI regulation is needed, India was the exception. People in India also expressed the most confidence in AI, with 75% of Indians surveyed saying they would place their trust in AI, followed by 67% of Chinese and 57% of South African respondents. The big picture: While India's AI and technology ecosystem currently lags behind other global leaders, its ministry views AI as a power boost for the digital economy and innovation. A relaxed approach to AI governance could encourage more talent and startups to enter the field, even while ethical concerns about the technology remain. 
| |
 A message from TRUERA LLM Testing and Monitoring – Join the Hands-On Workshop How can you ensure that your large language model is honest, harmless, and helpful? Join TruEra on May 11th and 18th as their President and Chief Scientist, Anupam Datta, provides a hands-on overview of how to analyze and improve the performance of your LLMs. During the two 60-minute online sessions, he will cover the following: - Quick introduction to LLMs - what are they and their history?
- Risks and challenges of LLMs: is your LLM honest, harmless, and helpful?
- Defining feedback functions to model LLM quality and performance
- How should you test and monitor LLMs?
- Live walk-throughs of how to identify, diagnose, and debug model issues in development and production
- Q&A with Professor Datta
Dates: 2-session series on May 11th and 18th Where: Virtual Event - Join from Anywhere Who should attend: Data scientists, machine learning engineers, MLOps managers, data analysts, statisticians, data and analytics managers Join for free
| |
AI-generated text detector GPTZero has launched a browser extension that can determine if online text is written by AI. The new tool, Origin, is designed to fact-check online media and will eventually be able to verify information sources, according to GPTZero. More: - Created by Edward Tian, a Princeton University student, GPTZero analyzes the randomness and "burstiness" of text to determine if it's written by AI.
- The company has now introduced Origin on the Google Chrome store. The browser extension enables users to highlight text on web pages to check if it's AI-generated.
- According to Tian, "We will fact-check each sentence on a website. If Origin identifies text as AI-generated, we will investigate the reference to determine whether it is true or false."
- GPTZero's 10-person team is now in talks with the BBC and industry executives, including former New York Times CEO Mark Thompson, on possible partnerships for AI detection and analysis to support journalism.
- The company also recently raised $3.5M in funding co-led by Uncork Capital and Neo Capital.
Zoom out: - GPTZero, an app that detects AI-generated text in essays and other documents, launched in January and has since drawn 1.2 million registered users.
- It reportedly has an accuracy rate of 99% for human text and 85% for AI text.
- ChatGPT creator OpenAI also introduced a tool, AI Text Classifier, that it says can distinguish between AI and human-written texts. The classifier was able to identify around 26% of content as "likely AI-written" in tests, making it not especially accurate.
BLOOMBERG
| |
AI automated voice ordering will soon roll out to some Carl's Jr. and Hardee's drive-thrus in the U.S. CKE Restaurants Holdings, the chains' parent company, is expanding the technology to more locations in partnership with AI technology providers Presto Automation, OpenCity, and Valyant AI. More: - The companies' voice automation tech will roll out to an undisclosed number of Carl's Jr. and Hardee's, which operate around 2,800 restaurants in 44 states and 1,000 outside the U.S.
- Pilot programs of AI technology from all three companies led to improved labor productivity at restaurants, according to CKE.
- Testing of Presto's automated voice ordering resulted in a "significant" increase in revenue at select locations, according to the firm.
- Presto said its Voice product was able to "upsell" customers more often, boosting the average check size at the restaurants.
Zoom out: USA TODAY
| |
AI Events and Conferences: - Google I/O: May 10, live-streamed and on-demand.
- AI and Big Data Expo North America: May 17-18, Santa Clara, California.
- AI World Congress 2023: June 7-8, London.
- Data Science Salon: Applying AI & Machine Learning to Media & Advertising: June 8, New York City and virtual.
- Big Data and Analytics Summit Canada: June 13-14, Toronto.
- Deep Learning World: June 18-22, Las Vegas.
- Ai4 2023: Aug. 7-9, Las Vegas.
- AI Hardware Summit: Sept. 12-14, Santa Clara, California.
- Edge AI Summit: Sept.12-14, Santa Clara, California.
- AI & Big Data Expo Global 2023: Nov. 30-Dec. 1, Olympia, London.
- The AI Summit New York: Dec. 6-7, New York City.
| |
QUICK HITS - Backed by Mark Cuban on Shark Tank, Collars & Co. makes polos and long-sleeve shirts that are lightweight, comfy, and have just the right amount of stretch. Use code INSIDE for 15% off.*
- General Assembly, a coding boot camp, has integrated ChatGPT into its course curriculum.
- Microsoft Advertising unveiled an "Ads for Chat" API that enables online services, publishers, and apps to select and provide ads via personalized chat experiences.
- Meta Platforms Inc. has recruited an Oslo-based team that previously worked on AI networking technology for Graphcore. The engineers were hired earlier this year and will aid in Meta's efforts to improve AI workload handling at its data centers.
- Get the latest Axios Pro report on what the current economic downturn means for your bottom line and how to plan for the future. Download for free.*
*This is a sponsored post.
| |
Upcoming events at Inside: - May 11 - Testing and monitoring large language models (LLMs) with TruEra (Register Here) *
- May 17 - Monthly Meditation guided by Nicholas Whitaker (Register Here)
- May 17 - Join Vanta's webinar to see Vendor Risk Management in action and learn about automating security reviews. (Register Here) *
- May 25 - Inside.com Book Club - Zero To One by Peter Thiel (Register Here)
*This is a sponsored listing.
| |
 |  | Beth is a writer and editor at Inside.com covering artificial intelligence and daily technology news. She has written the Inside AI newsletter since 2019. You can reach her at Beth.Duckett@inside.com. |   | | Editor | Aaron Crutchfield is based in the high desert of California. Over the last two decades, he has spent time writing and editing at various local newspapers and defense contractors in California. When he's not working, he can often be found looking at the latest memes with his kids or working on his 1962 and 1972 Fords. | |
Join us on May 11th and 18th for a hands-on overview of how to analyze and improve the performance of your LLMs. | |
Comments
Post a Comment