AI4Future: Top AI News (16-22 September)
Kate Shcheglova-Goldfinch, MSc MBA
CJBS Research Affiliate, AI lead EBRD/NBU, TCUK AML SWG co-chair and WCIB Freeman
This week’s news has prompted reflection on just how significant AI's role is for humanity, both in terms of its immense potential for good and the risks it poses. It has also underscored the importance of managing these technological transformations carefully. A new UN report, for instance, highlights the urgent need for global AI governance. The report, titled "AI Governance for Humanity", underscores AI’s great potential, while also warning of the serious risks it entails. According to the report’s authors, “The accelerated development of AI is concentrating power and wealth globally, with far-reaching geopolitical and geoeconomic implications.”
The UN is advocating for the establishment of new AI governance institutions, ensuring fair representation and coordination across key stakeholders to develop rules for a technology that transcends borders. Notably, the report reveals that while seven countries, including the US, UK, Germany, and Japan, are working together to coordinate AI regulation, 118 other nations are left out of these efforts. To appreciate the scale of AI’s impact on humanity, consider this: Precedence Research estimates that in 2024, the global AI market will be valued at $638 billion, growing to a staggering $3,680.47 billion by 2034.
The need for global AI management has only been heightened by recent breakthroughs, including the imminent arrival of ‘smart AI’ or AGI. Oracle has launched the first Zettascale cloud computing cluster, powered by NVIDIA’s Blackwell platform. To put this in perspective, Zetta is 1000 times Exa, meaning that a Zettascale cluster is 1000 times larger than an Exascale cluster - once considered enormous just a few years ago (remember, Zetta = 10^21). It’s believed the human brain operates at around 100 Petaflops, so a Zettaflop is about 10,000 times more powerful than the human brain. The future of AGI, it seems, isn’t too far off…
Another key trend is the increasing accessibility of custom AI. Mistral AI has launched a free tier and lowered the costs for using its language models. These updates aim to make their models more affordable and practical for developers and researchers, enhancing their cost-efficiency and performance in the AI market.
AI's adaptability to real-world applications is also on the rise. Retail giants Walmart and Amazon are gearing up to revolutionise the shopping experience through AI, while Lionsgate has granted AI research firm Runway access to a vast library of films and TV shows to train a new custom generative model. Meanwhile, investments in large-scale AI projects are surging. For instance, 11x.ai has raised $24 million to build an army of AI workers, while BlackRock and Microsoft have announced plans for a $30 billion fund dedicated to AI infrastructure.
In a notable twist, OpenAI is also seeing developments in its safety measures. CEO Sam Altman has stepped down from its internal safety board, which will now be headed by Carnegie Mellon professor Zico Kolter. The AI safety race is well and truly underway.
A round-up of this week’s key developments.
Walmart and Amazon drive retail transformation with AI
Walmart and?Amazon?are harnessing AI to drive retail transformation with new consumer experiences and enhanced operational efficiency. According to analytics firm?GlobalData, Walmart is focusing on augmented reality and AI-enhanced store management. Amazon, meanwhile, is leading advancements in customer personalisation and autonomous systems. GlobalData’s Disruptor Intelligence Center, utilising its Technology Foresights tool, has identified the strategic focus of these retail titans based on their patent filings. Walmart has submitted over 3,000 AI-related patents, with 20% of these in the last three years, indicating a swift evolution in its AI capabilities. In contrast, Amazon boasts more than 9,000 patents; half of which were filed during the same timeframe, underpinning its leadership in AI-driven retail innovations.
11x.ai raises $24M led by Benchmark to build AI digital employees
A startup that builds AI bots for process automation (aka automating end-to-end workflows), 11x.ai just raised a fresh $24 million Series A led by Benchmark. Founded in 2022, the company calls its AI agents “automated digital workers,” and like others in this field, its pitch is that its software can handle repetitive tasks, allowing human employees to focus on more strategic work.?
Sam Altman departs OpenAI’s safety committee
OpenAI CEO Sam Altman is leaving the internal commission OpenAI created in May to oversee “critical” safety decisions related to the company’s projects and operations. In a?blog post?today, OpenAI said the committee, the Safety and Security Committee, will become an “independent” board oversight group chaired by Carnegie Mellon professor?Zico Kolter, and including Quora CEO Adam D’Angelo, retired U.S. Army General Paul Nakasone, and ex-Sony EVP Nicole Seligman. All are existing members of OpenAI’s board of directors.
BlackRock and Microsoft plan $30bn fund to invest in AI infrastructure
Data centres require vast amounts of energy, creating bottlenecks in supply capacity. BlackRock is preparing to launch a $30 billion-plus artificial intelligence investment fund in partnership with tech giant Microsoft, aimed at building data centres and energy projects to meet the growing demands driven by AI. The financial partnership, spearheaded by BlackRock's new infrastructure investment arm, Global Infrastructure Partners, will be one of the largest investment vehicles ever created on Wall Street. Nvidia and Abu Dhabi-backed MGX are also set to join the effort.
California governor signs laws to protect actors against unauthorized use of AI
California Gov.?Gavin Newsom?signed on legislation aiming at protecting?Hollywood?actors and performers against unauthorized artificial intelligence that could be used to create digital clones of themselves without their consent. The new laws come as California legislators ramped up efforts this year to regulate the marquee industry that is increasingly affecting the daily lives of Americans but has had little to no oversight in the United States.
Mistral AI has taken a bold step in democratising access to its technology by introducing a free pricing plan and cutting the cost of its language models
Aimed at boosting availability for developers and researchers alike, these changes position Mistral’s models as both competitively priced and highly efficient, further strengthening its presence in the rapidly evolving AI market.
领英推荐
Lionsgate partners with AI firm to train generative model on film and TV library
Lionsgate has signed a deal with the artificial intelligence research firm Runway, allowing it access to the company’s large film and TV library to train a new generative model.
According to the?Wall Street Journal, the model will be “customized to Lionsgate’s proprietary portfolio” which includes hit franchises such as John Wick, Saw and The Hunger Games. The aim is to help film-makers and other creatives “augment their work” through the use of AI.
Oracle Offers First Zettascale Cloud Computing Cluster
Oracle today announced the first zettascale cloud computing clusters accelerated by the NVIDIA Blackwell platform.?Oracle Cloud Infrastructure?(OCI) is now taking orders for the largest AI supercomputer in the cloud - available with up to 131,072 NVIDIA Blackwell GPUs.
“We have one of the broadest AI infrastructure offerings and are supporting customers that are running some of the most demanding AI workloads in the cloud,” said Mahesh Thiagarajan, executive vice president, Oracle Cloud Infrastructure. “With Oracle’s distributed cloud, customers have the flexibility to deploy cloud and AI services wherever they choose while preserving the highest levels of data and AI sovereignty.”
U.N. says global rules for artificial intelligence are imperative
The United Nations is warning that artificial intelligence could inequitably advantage some parts of the world and leave others behind without a global strategy for how to oversee it.
In a?new report “Governing AI for Humanity” the United Nations proposed the formation of new AI governance institutions that are equitably represented and coordinated among participating groups to develop rules for a technology that has no borders. The report pointed out that seven countries, including the United States, Germany, UK and Japan, were working together to coordinate principles to govern AI, but 118 were not included in those efforts.?
Three Mile Island nuclear reactor to restart to power Microsoft AI operations
A nuclear reactor at the notorious Three Mile Island site in?Pennsylvania?is to be activated for the first time in five years after its owners, Constellation Energy, struck a deal to provide power to Microsoft’s proliferating artificial intelligence operations.
The plant was the location of the?most serious nuclear meltdown?and radiation leak in US history, in March 1979 when the loss of water coolant through a faulty valve caused the Unit 2 reactor to overheat. More than four decades later, the reactor is still in a decommissioning phase.
Meta, Ericsson, and many others sign open letter to EU over AI regulation
The open letter was coordinated by Meta, which managed to persuade Ericsson, SAP, Spotify and over 50 other companies to give it their seal of approval. It was published as an ad in the FT, which you can see a copy of below. It essentially argues that Europe’s excessive bureaucracy and characteristic lack of urgency on regulating AI means its companies risk falling behind other regions in seizing the opportunities presented by it.
LinkedIn suspends AI training using UK user data
LinkedIn has suspended the use of UK user data to train its artificial intelligence (AI) models after a regulator raised concerns. The career-focused social networking site, owned by Microsoft, quietly saw users around the world opted into their data being used to train its AI models. But the Information Commissioner's Office (ICO) said on Friday that it was "pleased" LinkedIn had confirmed that it had paused on using UK users' information. LinkedIn said it welcomes the chance to engage with the ICO further.
Insightful papers / books I have observed this week: