News

The Week in AI: OpenAI Adds Watermarks, Hugging Face Update, Altman Looks for Chip Investors, More

• By John K. Waters
• 02/09/2024

It's been another lively week in the AI space. This week's roundup includes the availability of a new messaging API from Hugging Face, a new SchoolAI partnership with Utah, a new API serving the DALL·E 3 model that includes C2PA metadata, the official rebranding of Google's Bard and launch of Gemini Advance, and Sam Altman's grand plan to get into the chip-making business.

OpenAI CEO Seeks Trillions to "Overhaul" Global Semiconductor Industry
Sam Altman says he wants to overhaul the semiconductor industry to accommodate the "massive-scale" infrastructure the evolving AI ecosystem will need in the future. The OpenAI CEO posted his goals on X:

"We believe the world needs more ai infrastructure--fab capacity, energy, datacenters, etc.--than people are currently planning to build. Building massive-scale AI infrastructure, and a resilient supply chain, is crucial to economic competitiveness. OpenAI will try to help!"

Altman is in talks with investors, The Wall Street Journal reported, including the United Arab Emirates. The project could require raising as much as $5 trillion to $7 trillion.

Hugging Face Intros Messages API
The open-source data science and machine learning platform community, Hugging Face, introduced the Messages API to provide OpenAI compatibility with Text Generation Inference (TGI) and Inference Endpoints. Beginning with version 1.4.0, TGI will provide an API compatible with the OpenAI Chat Completion API. The new Messages API allows customers and users to transition seamlessly from OpenAI models to open LLMs. The API can be directly used with OpenAI's client libraries or third-party tools, like LangChain or LlamaIndex. The community's website offers examples of inference endpoints on its blog page.

FCC Outlaws AI-generated Robocalls
The Federal Communications Commission (FCC) announced the unanimous adoption of a Declaratory Ruling that recognizes calls made with AI-generated voices are "artificial" under the Telephone Consumer Protection Act (TCPA). The ruling, which takes effect immediately, makes voice cloning technology used in common robocall scams targeting consumers illegal.  This would give State Attorneys General across the country new tools to go after bad actors behind these nefarious robocalls.

"Bad actors are using AI-generated voices in unsolicited robocalls to extort vulnerable family members, imitate celebrities, and misinform voters.  We’re putting the fraudsters behind these robocalls on notice," FCC Chairwoman Jessica Rosenworcel, said in a statement "State Attorneys General will now have new tools to crack down on these scams and ensure the public is protected from fraud and misinformation."

The Telephone Consumer Protection Act is the primary law the FCC uses to help limit junk calls.  restricts the making of telemarketing calls and the use of automatic telephone dialing systems and artificial or prerecorded voice messages.  Under FCC rules, it also requires telemarketers to obtain prior express written consent from consumers before robocalling them.  This Declaratory Ruling ensures AI-generated voices in calls are also held to those same standards.

K12 Brings Generative AI to Utah School District
K-12 generative AI platform provider SchoolAI announced a partnership with one of Utah's largest school districts to bring AI into classrooms for personalized one-on-one tutoring, guidance, and support. Jordan School District, located just outside Salt Lake City and home to 67 schools, 3,350 educators, and more than 57,800 students, will integrate AI into classrooms district-wide through SchoolAI's platform. "Teachers can take advantage of AI-powered lesson plans and offer a safe technology experience tailored to each student's individual needs," the company said.

The platform offers more than 1,000 activities with AI tutors, interactive games, simulations, well-being check-ins, and a library of grade- and subject-specific activities. Teachers get dashboards with real-time feedback and moderation, so they can track student progress and develop tailored learning plans.

Google Launches Gemini Advanced
Google officially rebranded its Bard AI model, Gemini, this week and launched Gemini Advanced, which gives users access to the Ultra 1.0 AI model, which the company says is far more capable at highly complex tasks, such as coding, logical reasoning, following nuanced instructions, and collaborating on creative projects. Gemini Advanced supports longer, more detailed conversations, the company says, and it also better understands the context from previous prompts.

For example:

• Gemini Advanced can be a personal tutor, creating step-by-step instructions, sample quizzes, or back-and-forth discussions tailored to the user's personal learning style.
• It can help with more advanced coding scenarios, serving as a sounding board for ideas and helping the developer evaluate different coding approaches.
• It can help digital creators go from idea to creation by generating fresh content, analyzing recent trends, and brainstorming improved ways to grow their audiences.

As Google adds new and exclusive features, Gemini Advanced users will have access to expanded multimodal capabilities, more interactive coding features, and deeper data analysis capabilities. Gemini Advanced is available today in more than 150 countries and territories in English, and Google plans to expand to more languages over time.

OpenAI Implements Image Authenticity Standards for C2PA
Images generated with ChatGPT on the web and the API serving the DALL·E 3 model will now include C2PA metadata. The C2PA (the Coalition for Content Provenance and Authenticity) aims to address "the prevalence of misleading information online through the development of technical standards for certifying the source and history or provenance of media content."

The update will be implemented for all users on mobile devices by February 12th. Sites like Content Credentials Verify can be used to determine if an image has been created with the DALL·E 3 model via OpenAI’s suite of tools. This verification will confirm whether the image was produced using OpenAI’s API or ChatGPT, except in cases where the metadata has been removed.

$10 million AI Math Olympiad Prize Gets New Advisors
XTX Markets, the London-based algorithmic trading company that launched a $10mn challenge fund last year known as the Artificial Intelligence Mathematical Olympiad Prize (AIMO Prize), announced the names of some prominent mathematicians and AI and machine learning specialists, who are joining the project as advisors. The group includes Timothy Gowers and Terence Tao, both winners of the Fields Medal, one of the most prestigious awards in mathematics; Po-Shen Loh, the former coach of the US IMO Team; Dan Roberts, an AI researcher at Sequoia Capital and MIT and a published expert in machine learning; and Geoff Smith, the former President of the International Mathematical Olympiad (IMO). The AIMO Advisory Committee will support the development of the AIMO Prize, including advising on appropriate protocols and technical aspects, and designing the various competitions and prizes.

XTX launched the fund to "spur the open development of AI models that can reason mathematically, leading to the creation of a publicly-shared AI model capable of winning a gold medal in the IMO."

The grand prize of $5mn will be awarded to the first publicly-shared AI model to enter an AIMO approved competition and perform at a standard equivalent to a gold medal in the IMO. The rest of the $5mn will be distributed in a series of prizes for publicly-shared AI models "that achieve key milestones towards the grand prize."