AI Weekly Digest - September 23 2024

News, views and innovations from the global artificial intelligence sector. LinkedIn newsletter readers can also sign up to an enhanced email edition of the AI Weekly Digest - published every Friday. You can subscribe for free.

US telco in OpenAI tie-up to build AI platform

Leading US telco T-Mobile US has announced a partnership with OpenAI to build an artificial intelligence platform to help it gain and retain customers. It said the new platform - IntentCX - will harvest data on customer interactions from the millions of T-Mobile subscribers using its recently launched T-Life app.

Altman eyes 'magical experiences' from T-Mobile US partnership

OpenAI CEO Sam Altman said he hopes the group can deliver "magical experiences" via its partnership with T-Mobile US, adding: “One of the many things we’re excited about for this new generation of models is what we can do for personalisation - for the individual user or customer - and what we can do for integration." T-Mobile CEO Mike Sievert said: “If you called us because you had a dropped call moments ago, AI can answer that call.” He went on: “Every one of those customers who left us left a separate individualised breadcrumb of data” about why they left... AI can ascertain what went wrong and why.”

Microsoft/BlackRock set to create $100bn Global AI Investment Partnership 

Microsoft is in talks with BlackRock about launching the Global AI Investment Partnership, a landmark artificial intelligence fund which is targeting $30bn in equity investment and a further $70bn in debt financing. The fund will focus on data centres and energy projects, and Microsoft president Brad Smith said: "The country and the world are going to need more capital investment to accelerate the development of the AI infrastructure needed. This kind of effort is an important step.

Lionsgate to use AI in film/TV production after Runway deal

US studio Lionsgate will start using generative artificial intelligence in the creation of its TV shows and films after signing a deal with AI startup Runway. The entertainment company behind The Hunger Games and Twilight will give Runway access to its content library in exchange for a new, custom AI model it can use in the editing and production process. Lionsgate chairman Michael Burns said it expects to save “millions and millions of dollars” from using the new model. He added: “We do a lot of action movies, so we blow a lot of things up and that is one of the things Runway does."

Reuters in multi-year deal with EZ Newswire

Reuters has agreed a multi-year deal with AI-enabled news distribution platform EZ Newswire. The start-up's founders Neel Shah and Caitlin Kelly said: “We are thrilled to collaborate with Reuters and look forward to further innovating in the years ahead to build the most powerful news and business intelligence platform."

YouTube pledges gradual, careful AI rollout as it unveils Veo features

CEO Neal Mohan told the Made on YouTube event that the platform will be "very gradual about the rollout" of AI tools in order to avoid unintended consequences. However, he was broadly supportive of the technology's long-term prospects, saying: “YouTube occupies a truly unique space. We get to work closely with cutting-edge technologies that are invented at YouTube, invented at sister organisations like Google DeepMind.”

FT's AccelerateAI 'a bid to explore opportunities and threats'

The Financial Times' creation of its AccelerateAI team earlier this year was an attempt to explore both opportunities and threats, Liz Lohn, director of product at FT said. She told the Future of Media Technology conference: "People are tired of talking about whether AI is going to kill us or save humanity." Lohn described the "underwhelming result" of training a large language model in the early stages of AccelerateAI, but said: "It was the best kind of failure, we’ve learned a lot from it." She went on: "Hallucinations are a bug in the media and a feature in LLMs," saying Gen AI will work better as an "inspire me" tool rather than as a content creator. She joked: "As soon as it works, no one calls it AI anymore."

OpenAI increases risk status of latest models

OpenAI has acknowledged that its latest models “meaningfully” increase the risk that AI will be misused to create biological weapons. After announcing its new o1 models last week, with their reported ability to reason, OpenAI’s system card - a tool to explain how the AI operates - said the new models had a “medium risk” for issues related to chemical, biological, radiological and nuclear weapons. This is the highest risk status OpenAI has ever given its models.

Scientist: Weapons risk reinforces need for AI regulation

Leading AI scientist Yoshua Bengio, a professor of computer science at the University of Montreal, said if OpenAI now represented “medium risk” for chemical and biological weapons “this only reinforces the importance and urgency” of legislation to regulate the sector.

Infosys chair: AI models to become commoditised

Infosys chairman Nandan Nilekani said he believes artificial intelligence models built by the likes of OpenAI will eventually become commoditised, with value creation flowing "to the application layer and the whole stack".

CEO: ChatGPT developer's annualised revenue rate hits $4bn

OpenAI chief executive Sam Altman told prospective investors the ChatGPT developer's annualised revenue rate recently hit $4bn, as he held talks about a $6.5bn fundraising that would value the group at $150bn. He also said it could change its not-for-profit corporate structure to make it more attractive to investors.

US to host international AI safety summit in November

The US government has convened a global safety summit on artificial intelligence on November 20 and 21, among ongoing Congress issues with regulating the new technology. Commerce secretary Gina Raimondo and secretary of state Anthony Blinken will host the inaugural meeting of the International Network of AI Safety Institutes in San Francisco, designed to "advance global cooperation toward the safe, secure, and trustworthy development of artificial intelligence".

'Godmother of AI' raises $230m for World Labs start-up

Leading artificial intelligence researcher Fei-Fei Li, widely known as the "godmother of AI," has raised $230m for start-up World Labs she and three colleagues founded. The operator's aim is to make AI technology which can understand how the three-dimensional physical world works. Li, a Stanford University professor, led AI at Google Cloud from 2017 to 2018, served on Twitter's board of directors and has been a key adviser to policymakers, including at the White House.

GSMA and 19 telcos launch Responsible AI Maturity Roadmap

The GSMA has launched a Responsible AI Maturity Roadmap alongside 19 telcos, including BT, Vodafone and Telefónica. The trade group's chairman José María Álvarez-Pallete López said: "The speed with which AI has now become a central part of tech and telecoms operations demonstrates its power and undoubted value but also the risks we must consider as an industry and the need to include ethics at the heart of AI to prevent its uncontrolled development. It is crucial for us all to ensure responsible guidelines for the use of AI are implemented now, and it is great to see the telecoms industry leading the way on this with the GSMA’s new roadmap.”

Report: ICO set to approve Meta harvesting data to 'supercharge AI technology'

The Information Commissioner's Office (ICO) is reportedly expected to approve plans from Meta to use millions of Britons' Facebook and Instagram data to 'supercharge its artificial intelligence technology'. A Telegraph report said the privacy watchdog is set to give the green light to for Meta to harvest 'billions of public posts and images from UK citizens'. Meta and the ICO declined to comment.

White House taskforce 'to address growing needs of AI infrastructure'

The White House has announced a new taskforce to deal with the growing needs of AI infrastructure. This followed a meeting between senior US officials and leading technology executives, including OpenAI CEO Sam Altman, Google senior executive Ruth Porat and Anthropic CEO Dario Amodei. Led by the National Economic Council and the National Security Council, the taskforce will coordinate policies to advance data centre development while weighing economic, national security, and environmental goals, a White House spokesperson said.

EU calls for AI operators to provide easier access to supercomputers

The EU has called for artificial intelligence operators to provide start-ups and researchers with easier access to supercomputers. These machines are designed for the large-scale training and development of general-purpose AI models, and for the operation running of emerging AI applications. Of the €2bn funding, half will come from Digital Europe and Horizon Europe, with the balance from participating EU member states. This will pay for new EU supercomputers, the upgrading of existing machines and creation of the AI factories.

Meta to reduce prominence of 'AI Info' tag

Meta is to give the 'AI Info' tag, which flags up content which has been manipulated using generative AI, a less prominent position on Instagram and Facebook.

Google search to flag AI-generated/edited images

Google announced it is to begin flagging AI-generated and AI-edited images surfaced by its Search service.