Author: IBL News

  • “Build Domain-Specific Language Models,” Said Coursera’s CEO in Davos [+Jan 18 Update]

    “Build Domain-Specific Language Models,” Said Coursera’s CEO in Davos [+Jan 18 Update]

    IBL News | New York

    Domain-specific applications using domain-specific data will unlock a lot of value and competitive advantage, said Coursera’s CEO Jeff Maggioncalda at Davos during the 54th annual meeting of the World Economic Forum.

    “Don’t try to build the next LLM for the world, try to build some domain-specific language models,” he stated. “It’ll be difficult to build businesses, unless you’re building foundation models.”

    “There is a huge untapped opportunity in building generative AI applications over building large language models (LLMs),” added Andrew NG, the founder and CEO of Landing AI and former CEO of Coursera.

    “There are so many more opportunities in applications of AI to healthcare, financial services, IT consulting, and so on. These have huge opportunities that relatively very few people are working on,” said Ng.
    .

    Update: On Thursday, January 18, Jeff Maggioncalda said that Coursera added a new user every minute on average for its artificial intelligence courses in 2023. The platform has more than 800 AI courses and saw more than 7.4 million enrollments last year.

    Every student on the platform gets access to a ChatGPT-like AI assistant called “Coach” that provides personalized tutoring.

    The chatbot uses OpenAI and Google’s Gemini’s LLMs.

    Maggioncalda said that the company does not plan to build or train its own models.“We’ll probably be fine tuning with proprietary data just on top of these large language models.”

    Coursera has also used the technology to translate about 4,000 courses in different languages and plans to ramp up hiring for AI roles this year.

  • Indian Company Upgrad Education to Acquire Udacity for $80 Million, According to Local Media

    Indian Company Upgrad Education to Acquire Udacity for $80 Million, According to Local Media

    IBL News | New York

    Upgrad Education, led by Indian entrepreneur Ronnie Screwvala, is in the advanced stages of acquiring this year for $80 million the Silicon Valley online education platform Udacity, one of the three big MOOC platforms, along with Coursera and edX/2U, according to The Economic Times.

    Sebastian Thrun-founded Udacity was last valued at $1 billion in 2015 after it raised $105 million from investors like Germany’s Bertelsmann and Scotland’s Baillie Gifford. VC Andreessen Horowitz (a16z) is also one of its investors.

    Upgrad founder Ronnie Screwvala [in the picture], who owns of 22.4% in Upgrad, has pitched to external investors for about $50-60 million.

    According to the same source, the Mumbai-based Upgrad is trying to raise up to $100 million, of which about 80% will be used for financing the deal.

    If this funding round is completed, the valuation for Upgrad will see a jump from its last assessment of $2.5 billion.

    Upgrad has raised $365 million since its inception in 2015.

    California-based Udacity offers courses across data engineering, business analytics, artificial intelligence, data science, product management and others. Udacity has three key verticals: individuals, government, and enterprise. The enterprise business is the largest revenue driver for the US-based firm.
    .

  • An AI-powered Diagnostic Tool Used by Kaiser Permanente Gets a High Valuation

    An AI-powered Diagnostic Tool Used by Kaiser Permanente Gets a High Valuation

    IBL News | New York

    Nabla, a French start-up that develops AI copilots for doctors and medical staff, reached a valuation of $180 million following a fundraising of $24 million in Series B funding.

    The round, led by Cathay Innovation, with participation from ZEBOX Ventures, came a few months after Nabla signed a large-scale partnership with U.S. healthcare giant Kaiser Permanente, and now it is actively used by thousands of doctors.

    Nabla’s main copilot has been described by the company as “a silent work partner that sits in the corner of the room, takes notes, and writes medical reports.” It can help doctors save time on admin work so that they can be more focused on patients.

    Essentially, when a physician starts a consultation, he hits the start button in Nabla’s interface, forgetting about their computer. It uses real-time speech-to-text technology to turn the conversation into a written transcript. Nabla uses a combination of a speech-to-text API from Microsoft Azure and its technology based on the open-source Whisper model.

    The tool works with both in-person consultations and telehealth appointments.

    It uses an LLM refined with medical data and health-related conversations to identify the important data points in the consultation — medical vitals, drug names, pathologies, etc.

    At the end of the consultation, the system generates a thorough medical report with a summary of the consultation, prescriptions, and follow-up appointment letters.

    The report can be personalized by the doctor, with instructions to be more concise or verbose, along with notes that follow the Subjective, Objective, Assessment, and Plan (SOAP) pattern, widely used in the U.S.

    Physicians have a final say as they can edit reports before they are filed in their electronic health record system (EHR).

    Nabla is currently available as a web app or a Google Chrome extension.

    The Paris-based company ensured TechCrunch that it is on track to process more than 3 million consultations per year in three languages.
    .

  • Coursera Translated 4,200 Courses Into 17 Languages with AI

    Coursera Translated 4,200 Courses Into 17 Languages with AI

    IBL News | New York

    Coursera, with over 129 million registered learners, has translated 4,200 courses into 17 languages with Generative AI, the company’s CEO, Jeff Maggioncalda, said to Forbes.

    Coursera has also deployed a personalized tutor chatbot named ‘Coach.’ Students can summarize learning activities, ask for help on a concept, or create practice problems.

    Coach emphasizes learning, but it doesn’t give users the answer, especially during end-of-lesson quizzes.

    Separately, the company is utilizing AI to help its 300+ university and industry partners build courses and develop content. The tool allows to generate outlines, write learning objectives, and compile lessons into new courses.

    Maggioncalda says that the company is aiming into an era of having the most adaptive, personalized, and effective learning experience that a student can have.
    .

     

  • The 2024 Open edX Annual Conference Will Take Place in Early July in South Africa

    The 2024 Open edX Annual Conference Will Take Place in Early July in South Africa

    IBL News | New York

    The Open edX organization will host its 2024 annual conference in Cape Town, South Africa, between July 2nd and July 5th, 2024.

    The event will take place at Stellenbosch University.

    The organizers — the Center for Reimagining Learning (or “tCRIL”), the MIT and Harvard non-profit organization — announced that the conference will showcase innovative technologies and use cases, advancements in instructional design, and methods for operating and extending the Open edX platform.

    In addition, breakthrough applications in generative AI will be featured.

    The call for proposals will end on January 17th. Registration is open, along with sponsorship options.

  • OpenAI Launched Its GPT Store, Which Features Custom Versions of ChatGPT

    OpenAI Launched Its GPT Store, Which Features Custom Versions of ChatGPT

    IBL News | New York

    OpenAI introduced yesterday its GPT Store, with millions of custom versions of ChatGPT only open to paid users under ChatGPT Plus, Team, and Enterprise premium plans.

    The store features over three million GPTs, including trending chatbots on the community leaderboard, with categories like DALL·E, writing, research, programming, education, and lifestyle. IBL News has its version in this store.

    OpenAI is highlighting these GPTs: 

    • Personalized trail recommendations from AllTrails
    • Search and synthesize results from 200M academic papers with Consensus
    • Expand your coding skills with Khan Academy’s Code Tutor
    • Design presentations or social posts with Canva
    • Find your next read with Books
    • Learn math and science anytime, anywhere with the CK-12 Flexi AI tutor

    OpenAI announced that it will launch a revenue program for GPT creators in the U.S. based on user engagement during the first quarter of this year.

    In addition, OpenAI launched a new ChatGPT Team plan for teams of all sizes, priced at $25/month/user (annual) or $30/month (monthly), offering a shared workspace and user management tools.

    OpenAI started rolling out ChatGPT personalization, allowing GPT to carry over memory between chats and improve over time.

  • AI-Created Social Avatars May Threat the Business of Human Influencers

    AI-Created Social Avatars May Threat the Business of Human Influencers

    IBL News | New York

    Hundreds of AI-generated influencers have broken into the $21 billion content creator economy, threatening human characters.

    Brands have been quick to engage with virtual influencers as a new way to attract attention while reducing costs. In addition, they have total control versus a real person who comes with potential controversy and their own opinions.

    These virtual avatars, hyper-realistic created with AI, can be followed by large audiences, as the Financial Times reported.

    For example, pink-haired, digital Aitana López is followed by over 200,000 people. It was created by the Barcelona-based agency The Clueless, whose creations [in the picture] have been criticized for being overly sexualized.

    Brands such as Victoria’s Secret have paid $1,000 a post to promote their products.

    Other virtual influencers like Lil Miquela achieved an audience of 3 million followers and have worked with Burberry, Prada, and Givenchy.

    The rise of these virtual influencers has been accelerated by the fact that no rule obliges creators to declare that they are generated by AI. India is the only country that forces virtual influencers to reveal their AI origins.

    H&M said that its virtual influencer Kuki has reached 11 times more people and resulted in a 91 percent decrease in cost per person compared with a traditional ad.
    .

  • OpenAI Fires Back at The New York Times’ Copyright Lawsuit: “The NYT Manipulated Its Models”

    OpenAI Fires Back at The New York Times’ Copyright Lawsuit: “The NYT Manipulated Its Models”

    IBL News | New York

    OpenAI responded in a blog post yesterday to The New York Times about its legal action, saying that the paper “manipulated” its models, “is not telling the full story”, and its lawsuit is “without merit.”

    • “Training is fair use, but we provide an opt-out because it’s the right thing to do,” OpenAI said.

    • “The principle that training AI models is permitted as fair use is supported by a wide range of academics, library associations, civil society groups, startups, leading US companies, creators, authors, and others that recently submitted comments to the US Copyright Office.”

    • “We have led the AI industry in providing a simple opt-out process for publishers (which The New York Times adopted in August 2023) to prevent our tools from accessing their sites.”

    • “The New York Times is not telling the full story. Our discussions with them appeared to be progressing constructively through our last communication on December 19. (…) Their lawsuit on December 27 came as a surprise and disappointment to us.”

    • “It seems The New York Times intentionally manipulated prompts, often including lengthy excerpts of articles, in order to get our model to regurgitate.”

    • “We regard The New York Times’ lawsuit to be without merit.”

  • AI-powered TV Anchors Rapidly Spread in Asia

    AI-powered TV Anchors Rapidly Spread in Asia

    IBL News | New York

    AI-powered TV anchors are rapidly spreading in Asia. One of the most popular is Lisa, the Indian broadcaster Odisha TV’s anchor [in the picture above].

    This robot delivers news bulletins, horoscopes, and weather and sports updates.

    Also, Delhi-based India Today Group features Sana, an AI agent that presents news and the weather in English, Hindi, and Bangla, and co-anchors programs with other journalists in 75 languages.

    In the southern state of Karnataka, Power TV is using Soundarya, who introduced herself as a “robot anchor.”

    This growing tribe of AI anchors also collects and analyzes data, converting it into usable and actionable information.

    Production managers say they save costs, allow channels to deliver news in a country with 22 official languages, and crunch vast amounts of data at phenomenal speed.

    Critics say the technology is undermining media credibility as they deliver news with a non-human monotonous voice and no hand gestures.

    China, Malaysia, and Taiwan broadcasters also use virtual presenters.
    .

  • Outgoing CEO at SNHI Anticipates that AI Will Be Deeply Disruptive

    Outgoing CEO at SNHI Anticipates that AI Will Be Deeply Disruptive

    IBL News | New York

    Paul J. LeBlanc, President and CEO of Southern New Hampshire University (SNHU), who will step down on June 30, 2024, anticipates that the next disruptive wave will be AI. “AI means universities must change dramatically,” he said to The Boston Globe last week.

    Paul J. LeBlanc plans to work with a team of six researchers to study emerging AI trends, impacts on education, and opportunities to innovate. He persuaded researcher George Siemens to leave his post and join the team as chief scientist, he revealed to EdSurge.

    The group is working on a learning platform as well as setting up to build a massive data consortium in higher-ed, scheduled to be launched in April 2024 at the ASU-GSV conference. 

    • “I think AI will do to knowledge work — or white-collar jobs if you prefer — what automation did to blue-collar jobs,” he says. “It’s going to be deeply disruptive and displace a lot of people, and we’re going to have to adapt around that.”

    • “There are also bigger questions about what jobs will go away and what jobs will be created, which influences the fields of study schools will offer.”

    • “In the not-so-distant future, professionals will be judged less by what they know and more by what they can do with what they know, LeBlanc said.”

    Paul J. LeBlanc led SNHU through major growth in his two decades, transforming a school with 2,500 students in 2003 to more than 225,000 learners, the vast majority of them online, that it serves today.

    The Board of Trustees unanimously voted to appoint Lisa Marsh Ryerson as President for a two-year term, effective July 1, 2024.
    .