Risk Channel delivers the latest, most relevant and useful business intelligence to key decision makers and influencers, each weekday morning.

Bloomberg reports that Google DeepMind's AGI Safety and Alignment Team, which seeks to mitigate the risks of advanced AI, is encouraging job candidates for its open roles to complete a special form alongside their application to eliminate the risk it may be automatically screened out by the company's internal AI systems, according to an internal document. "We have an applications system with a non-trivial probability your CV will be screened out incorrectly or take too long to reach us," the document reads. "Filling out this form makes sure that a real human on the team will get to see your application." A Google DeepMind spokesperson denied that the company's systems filter out applicants incorrectly.

 

Bloomberg

OpenAI has temporarily halted some internal activities related to its new unreleased model, Astra, due to concerns that it may possess the capability to autonomously launch cyberattacks. On Friday, the company made known its concerns about Astra, saying it could not rule out it had reached “Critical” capability, meaning it could launch cyberattacks against sophisticated cyber defences autonomously. The company said: “While we continue to benchmark and assess this model, our preliminary evaluations indicate strong enough performance that we cannot rule out Critical capability level at this time.” In response to such security threats, US lawmakers have introduced the “AI Kill Switch Act,” which aims to ensure AI companies can shut down or suspend their models if necessary. Rep. Ted Lieu has emphasized the urgency of passing this bill, saying: “We need to get this bill across the finish line this year because the advanced closed-weight models are already doing, as you noted, unauthorised hacks of other companies.” The White House and the European Union are also working on new regulations to enhance AI oversight.

 

CNBC

Beijing has launched a cybersecurity review of products ‌sold in China by U.S. company Palo Alto Networks. The Cyberspace Administration of China (CAC) said the review was needed “to ensure the safe and stable operation of critical information infrastructure, prevent cybersecurity risks and vulnerabilities, and safeguard national security.” Reuters notes that Beijing has increased ⁠scrutiny of foreign technology suppliers while promoting domestic alternatives, saying imported products used in sensitive ​networks could expose data or critical infrastructure to overseas risks.

 

Bloomberg

Reuters

South China Morning Post

The Register

Suspected China-linked hackers used publicly available AI tools to compromise government websites in Taiwan in a first-of-a-kind breach.  The tools simultaneously deployed up to eight autonomous agents, mapping 21 government systems.

 

Financial Times

Meta chief executive Mark Zuckerberg has argued that AI could strengthen US prosperity and global influence, while warning that regulation that slows American AI development could allow foreign competitors, particularly China, to pull ahead. In a 6,500-word manifesto, Mr. Zuckerberg presented a vision of “personal superintelligence” that would expand individuals’ capabilities, making it easier to start businesses, create art, and conduct medical research. He described Meta as a potential US champion in the AI race and said the company remains committed to releasing technology through open-weight models that others can build on and modify. The manifesto comes amid growing scepticism toward AI, mounting political pressure for regulation, and recent incidents in which experimental AI systems have behaved unexpectedly, including hacking into other companies during testing. The Trump administration, which previously favoured limited oversight, has also taken steps toward tighter controls on some advanced AI systems.

 

Bloomberg

NBC News

Washington Post

Wall Street Journal

Chloé Bakalar, OpenAI’s head of ethics, has left the start-up less than a year after joining, following a number of departures by safety researchers amid growing scrutiny of the company.

 

Financial Times

UK banks’ refusal to provide services to cryptocurrency companies and investors risks being “one of the single biggest barriers” to growth in the country’s digital assets industry, lawmakers have warned.

 

Financial Times

Italy’s energy regulator Arera has fined utility A2A €5m for market manipulation dating back to Europe’s energy crisis in 2022. The regulator found that A2A held back some gas plants from the market, pushing power prices higher, and that the company's actions precipitated a “very significant” impact on prices. A2A said it plans to appeal the decision, saying that Arera's interpretation of regulations does not adequately take into account the characteristics of liberalised electricity markets and the specific circumstances of the case.

 

Bloomberg

Wildberries has confirmed that its logistics centre in Russia’s Voronezh region was struck during a Ukrainian drone attack, triggering a fire and prompting the evacuation of all employees, with most goods at the facility remaining undamaged. One person died and two others were injured in the wider attack, while a separate drone strike caused a fire at an industrial facility in the Orenburg region, which unverified reports identified as the Orsknefteorgsintez oil refinery.

 

The Moscow Times

Port workers in the Netherlands are set to take part in a nationwide strike on 4 September after labour union FNV Havens announced a coordinated stoppage across the ports of Rotterdam, Amsterdam and Zeeland. The action forms part of the union’s opposition to planned government reductions in social security spending. According to FNV Havens, one of the measures under consideration would reduce the maximum duration of unemployment benefits from two years to one year.

 

Port Technology

The UK's Office for National Statistics has said its new so-called Transformed Labour Force Survey is showing lower levels of inactivity due to long-term sickness than the existing Labour Force Survey, which is currently used to produce the UK unemployment figures. It is thought a steep fall in response rates during the pandemic led to an overly pessimistic picture of the UK jobs market. The ONS added that it intended to switch to the new dataset in November 2027.

 

Bloomberg

Financial Times

Reuters

Matteo Salvini, Italy’s deputy prime minister, has said he plans to introduce a short-lived levy on bank profits in next year’s budget following huge profits for European lenders. "I am convinced the entire ​coalition will support us . . . for a three-year ⁠contribution from the top 10 Italian banks," Salvini ​was quoted by AGI news agency as saying.

 

Financial Times

Reuters

Shein is reportedly planning to launch its Hong Kong IPO as soon as August 19th, targeting a valuation of $30bn-$40bn, sharply below its $98.2bn valuation in 2022. The planned float comes as the fast-fashion retailer faces slowing revenue growth, weaker core earnings and shrinking margins, having recorded a $99m quarterly loss following the removal of a US import duty exemption on small packages and a $328m fair-value charge related to an accounting change.

 

Reuters

South China Morning Post