AI Weekly Digest - April 22 2024

Here is PA Media's weekly need-to-know round-up of key news from the artificial intelligence sector. LinkedIn newsletter readers can also sign up for free to an enhanced email edition of the AI Weekly Digest - published every Friday. You can subscribe for free and benefit from:

• A full rundown of the week's most important AI news, curated by PA Media's editorial team
• Earlier access to the digest with our email every Friday
• Additional content with our AI Tools Spotlight highlighting new technologies and tools coming to market
• Regular longer reads from PA's own journalists with our AI Insights

Microsoft looks beyond OpenAI with G42 stake

Microsoft has agreed to pay $1.5bn for a minority stake in Abu Dhabi artificial intelligence group G42. The tech giant's president Brad Smith will take a seat on the start-up's board as part of the deal. Smith said: “Microsoft’s large investment is not something we do without a lot of thought. And this decision reflects confidence by our company in the UAE as a country, in G42 as a company, and in Peng (Xiao) as its CEO.” 

OpenAI takes on Microsoft with business pitches

Sources told Reuters that OpenAI made pitches to hundreds of businesses in London, New York and San Francisco earlier this month, regularly pitting its services against key financial backer Microsoft. The talks were hosted by CEO Sam Altman and other senior executives.

ChatGPT pioneer plans big push into Asia

Meanwhile, OpenAI has opened its first office in Asia, establishing a team in Tokyo targeting contracts with Japanese companies. CEO Sam Altman said: "This is just the first step in what I hope will be a long-term partnership with the people of Japan, government leaders, businesses and research institutions."

Google backs referral traffic as it launches search trial

Google is trialling search answers written by artificial intelligence for a small number of logged-in users in the UK, although spokeswoman Hema Budaraju said it remains a "priority" to generate referral traffic for creators, and added AI-powered results are "actually showing more links to a wider range of sources".

Campbell Brown joins AI media marketplace

Meta's former head of news Campbell Brown has joined TollBit, a US start-up building a marketplace connecting AI providers and media companies. Axios said it understands she had held talks with groups including OpenAI before taking up her new role.

DeepMind says Google's AI spend to top $100bn

DeepMind chief executive Demis Hassabis said parent Google expects to invest more than $100bn in artificial intelligence. He also said his group agreed to be acquired by Google in part because it has superior computing power to rivals: "That is one of the reasons we teamed up with Google back in 2014, is we knew that in order to get to (artificial general intelligence) we would need a lot of compute. That is what’s transpired, and Google had and still has the most computers.”

Amazon CEO says impact of AI will 'astound us all'

Amazon CEO Andy Jassy said generative artificial intelligence "may be the largest technology transformation since the cloud (which itself, is still in the early stages), and perhaps since the internet", adding in his annual letter to shareholders the "amount of societal and business benefit from the solutions that will be possible will astound us all". He also said: "We’re optimistic that much of this world-changing AI will be built on top of AWS."

Hass warns on power-hungry AI models

Arm Holdings CEO Rene Haas said training AI models could push global electricity infrastructure to its limits, forecasting data centres will consume more power than India by 2030. He said: "There needs to be broad breakthroughs. Any piece of efficiency matter."

No 10 'switches stance on artificial intelligence regulation'

Sources told the FT that the government is drawing up new legislation to regulate artificial intelligence, six months after PM Rishi Sunak asked: “How can we write laws that make sense for something that we don’t yet fully understand?” A person briefed on the government's plans said: “Officials are exploring moving on regulation for the most powerful AI models."

Vestager says AI poses 'individual existential risks'

EU competition commissioner Margrethe Vestager said humanity has never before been "confronted with a technology with so much power and no defined purpose", telling a Princeton symposium on AI that the technology raises "individual existential risks".

Almost half of smartphones to be AI-enabled

Counterpoint Research said it expects generative AI-capable handsets to account for almost half of smartphone shipments in 2027, up from 11% this year. It said: “We expect such devices to have multimodal capabilities, allowing them to process text, image, voice and other inputs to generate a variety of output and enable a user experience that is fluid and seamless. We expect the hardware specifications of such devices will likely evolve as the technology advances. But at present, a device should have hardware capabilities that are comparable to or exceed the performance of current flagship smartphones to effectively run generative AI models.” 

Meta says Llama 3 will be less 'sanctimonious'

Facebook owner Meta has unveiled a less "sanctimonious" version of its AI model, Llama 3, which the company said has “vastly improved capabilities”. President of global affairs Nick Clegg said the update came amid criticisms that the AI model refused to answer “innocent and innocuous” questions. He told the FT: “People, I think, found [Llama 2] a little bit sanctimonious, a little bit reluctant to engage. So we’ve really, really tried to work hard to reduce what we call ‘false refusals’ to make the underlying models more helpful and more responsive.”