Latest news with #AIAct
Euronews
a day ago
- Business
- Euronews
Irish privacy watchdog seeks extra funding to handle expanded role
The Irish Data Protection Commission (DPC) has said it needs more money to carry out additional tasks it now handles, including oversight of the EU AI Act. The AI Act – which regulates the technology according to the risk it poses to society – has already entered into force, but as of 2 August all member states need to appoint an oversight authority to ensure companies' compliance with the rules. In its annual report published Thursday, the DPC said that 'in light of new responsibilities and a significantly additional workload for the DPC as a result of the AI Act and other digital regulations [...] it is critical that we continue to receive funding increases enabling the expansion of our workforce.' 'The Government's continuing support will be critical to the DPC's ability to meet its EU wide responsibilities and the delivery of effective regulation in support of the digital economy,' it added. This year, the Irish already dealt with several AI questions, stemming from the launch of chatbot tools such as X's Grok and Meta AI. As the lead authority for Meta, it ordered the company to halt the tool last year due to concerns about the use of personal data of users of Facebook and Instagram to train its large language models (LLMs). Euronews reported in May that – with months to go until the deadline – in at least half of the 27 member states, it remains unclear which authority will be nominated as AI oversight body. In addition, countries need to adopt an implementing law that sets out penalties and that empower their watchdogs. Not all of them have yet done so. The Irish watchdog is currently overseeing the implementation of the General Data Protection Regulation (GDPR) in Ireland, including those of the global big tech companies that registered their EU headquarters in Dublin. It received some 11,091 new cases and resolved 10,510 ones, the report said. It gathered a total of €652 million in fines. Its staff increased from 213 in early 2024 to 251 as of 1 January. The use of a powerful American 'bunker bomb' against Iran's nuclear facilities continues to weigh over the country's conflict with Israel. CNN reported that US President Donald Trump is warming to the idea of using military assets to strike Iranian nuclear facilities and that he is 'not much in the mood to negotiate with Iran'. One of the potential weapons that the US military has at its disposal, the GBU-57 A/B Massive Ordnance Penetrator (MOP), could reach one of Iran's key nuclear sites, which is buried deep in a mountain. Iran's deputy foreign minister, Kazem Gharibabadi, said on state TV on Thursday that Iran 'will have to use its tools to both teach a lesson to aggressors and defend its national security and national interests,' if the US decides to intervene in the conflict. The GBU-57 MOP is a weapons system that is designed to attack 'hard and deeply buried targets such as bunkers and tunnels,' according to a 2024 US Air Force report. An archived fact sheet from the US Defense Threat Reduction Agency (DTRA) says it is a '30,000 pound (13.6 tonnes) class penetrator bomb,' that is approximately 20.5 feet long and has a diameter of 31.5 inches. The DTRA said the weapon can carry over 5,300 pounds of explosive material and delivers '10 times' the explosive power of its predecessor, the BLU-109. The Air Force Research Laboratory (AFRL) first contracted Boeing for the MOPs in 2004 with an initial contract worth $30 million (€26.2 million). The weapon was developed in three stages, the DTRA continued. The Air Force received the first bombs for testing in 2011, according to a report from the San Francisco Chronicle. The ordnance has been tested at the White Sands Missile Range, a military base in the US state of New Mexico. The Fordo Fuel Enrichment Plant (FFEP) is buried deep under a mountain near Qom, Iran, and is believed to be one of Iran's key uranium enrichment facilities for it's nuclear weapons programme, according to Heather Williams, director of the project on nuclear issues at US-based Center for Strategic & International Studies (CSIS). Rafael Mariano Grossi, the IAEA's director general, said on June 16 that there is so far no damage to the Fordo site, unlike the one at Natanz, which saw its pilot fuel enrichment plant destroyed. A recent report from the International Atomic Energy Agency (IAEA) estimates that 166.6 kilograms of 60 per cent enriched uranium of the total 408.6 kilograms that Iran has at that enrichment percentage was produced at Fordow. Uranium needs to be enriched to roughly 90 per cent to become a nuclear weapon but experts have previously said that Iran could create weapons at 60 per cent enrichment. Israel lacks the ordnance to take out Fordow on its own, but 'multiple strikes' from the US GBU-57 that are mounted on B-2 bombers 'could destroy the facility,' Williams said. If President Trump decides to use the GBU-57, it could 'constitute direct support for Israel and have the potential to escalate and drag the US into another war in the region,' Williams wrote. There are other options that Israel could use to get to Fordo, Williams continued, like continual strikes at the facility by GPU-28 or BLU-109 weapons that can burrow into a target and could target above-ground entrances or exits to the facility.
The Sun
4 days ago
- Politics
- The Sun
Malaysia needs comprehensive AI-focused laws
BANGI: Malaysia must introduce dedicated legislation to regulate artificial intelligence (AI) and mitigate its associated risks, said Minister in the Prime Minister's Department (Law and Institutional Reform) Datuk Seri Azalina Othman Said yesterday. Speaking after her ministry's monthly assembly and the launch of the National Legal Academy, Azalina confirmed she had formally written to Digital Minister Gobind Singh Deo to initiate discussions on developing AI-specific laws. 'I've proposed a meeting between the Legal Affairs Division and the Digital Ministry to explore a legislative framework for AI,' she said. Azalina stressed that AI operates on a fundamentally different paradigm from conventional technologies and cannot be governed under existing legal frameworks. She pointed to the European Union's AI Act – the world's first comprehensive AI legislation – as a model Malaysia could study and adapt to local needs. 'We must look at global frameworks like the EU's AI Act, which categorises risks and defines responsibilities, to see how they can inform Malaysia's approach,' she said. She also highlighted broader efforts to digitise the judicial system, including the use of voice-to-text transcription and online filing, noting that such systems must extend beyond Putrajaya to district courts nationwide. Gobind had earlier announced that a detailed proposal on Malaysia's AI regulatory roadmap, currently being finalised by the National Artificial Intelligence Office, would be ready by the end of June. The roadmap may include new legislation, regulatory rules or standard-setting frameworks. Azalina also called for stronger corporate accountability laws, citing the recent fatal bus crash in Gerik that claimed 15 lives. 'We need serious policy discussions around corporate negligence and potential laws on corporate manslaughter, especially when public safety is compromised,' she said. Separately, she confirmed that the Online Safety Act 2024 – which has been gazetted – is awaiting its enforcement date. The Act will come into effect once Communications Minister Datuk Fahmi Fadzil signs the enforcement order. Once in force, the law will allow for immediate regulatory action against harmful online content and introduce a statutory duty of care on platform providers. An online safety committee will also be formed to define and classify online harms, although child-related matters will remain governed under the Child Act 2001. Azalina confirmed she had formally written to Digital Minister Gobind Singh Deo to initiate discussions on developing AI-specific laws. – Amirul Syafiq/theSun
The Sun
4 days ago
- Politics
- The Sun
Azalina urges AI laws, cites EU model as reference
BANGI: Malaysia must introduce dedicated legislation to regulate artificial intelligence (AI) and mitigate its associated risks, said Minister in the Prime Minister's Department (Law and Institutional Reform) Datuk Seri Azalina Othman Said yesterday. Speaking after her ministry's monthly assembly and the launch of the National Legal Academy, Azalina confirmed she had formally written to Digital Minister Gobind Singh Deo to initiate discussions on developing AI-specific laws. 'I've proposed a meeting between the Legal Affairs Division and the Digital Ministry to explore a legislative framework for AI,' she said. Azalina stressed that AI operates on a fundamentally different paradigm from conventional technologies and cannot be governed under existing legal frameworks. She pointed to the European Union's AI Act – the world's first comprehensive AI legislation – as a model Malaysia could study and adapt to local needs. 'We must look at global frameworks like the EU's AI Act, which categorises risks and defines responsibilities, to see how they can inform Malaysia's approach,' she said. She also highlighted broader efforts to digitise the judicial system, including the use of voice-to-text transcription and online filing, noting that such systems must extend beyond Putrajaya to district courts nationwide. Gobind had earlier announced that a detailed proposal on Malaysia's AI regulatory roadmap, currently being finalised by the National Artificial Intelligence Office, would be ready by the end of June. The roadmap may include new legislation, regulatory rules or standard-setting frameworks. Azalina also called for stronger corporate accountability laws, citing the recent fatal bus crash in Gerik that claimed 15 lives. 'We need serious policy discussions around corporate negligence and potential laws on corporate manslaughter, especially when public safety is compromised,' she said. Separately, she confirmed that the Online Safety Act 2024 – which has been gazetted – is awaiting its enforcement date. The Act will come into effect once Communications Minister Datuk Fahmi Fadzil signs the enforcement order. Once in force, the law will allow for immediate regulatory action against harmful online content and introduce a statutory duty of care on platform providers. An online safety committee will also be formed to define and classify online harms, although child-related matters will remain governed under the Child Act 2001. Azalina confirmed she had formally written to Digital Minister Gobind Singh Deo to initiate discussions on developing AI-specific laws. – Amirul Syafiq/theSun
Sinar Daily
4 days ago
- Business
- Sinar Daily
Malaysia mulls drafting AI-focused laws
Malaysia has no specific laws focused on AI because, unlike traditional technologies, AI operates on an entirely different platform. 16 Jun 2025 03:13pm Minister in the Prime Minister's Department (Law and Institutional Reform) Datuk Seri Azalina Othman Said. Bernama FILE PIX BANGI - Malaysia is studying to develop artificial intelligence (AI) legislation to address legal complexities in the digital age, said Minister in the Prime Minister's Department (Law and Institutional Reform) Datuk Seri Azalina Othman Said. Azalina said she has written to Digital Minister Gobind Singh Deo to propose a meeting between the Legal Affairs Division (BHEUU) and the Digital Ministry to initiate a discussion on drafting new AI laws. "I have written to the Digital Minister Gobind about looking into more AI (specific) laws for our country. There should be a meeting between BHEUU and the Digital Ministry," she said. Azalina was speaking at a press conference after BHEUU's monthly assembly and the launch of the National Legal Academy (APN) here today. Minister in the Prime Minister's Department (Law and Institutional Reform) Datuk Seri Azalina Othman Said. Bernama FILE PIX She said Malaysia has no specific laws focused on AI because, unlike traditional technologies, AI operates on an entirely different platform. Azalina said that so far, the only comprehensive AI legislation in the world is the European Union's AI Act, which sets out clear definitions, responsibilities and categories of AI risks. Azalina also highlighted Malaysia's progress in judicial digitalisation, including initiatives like voice-to-text systems and online filings. "We want the court to be more efficient. So the court is taking in more digital documents, but we want more to be done. We want it to apply not only in Putrajaya but also elsewhere, especially at district courts," she added. - BERNAMA
Phone Arena
4 days ago
- Phone Arena
Microsoft rolls out the Copilot feature that can watch your every move
A couple of months ago, we talked about Microsoft's new Copilot Vision feature for Windows and now, it is officially here. This new AI-powered tool gives Copilot the ability to actually see what is on your screen and interact with it in real time. Right now, Copilot Vision with Highlights is rolling out to users in the US, with Microsoft planning to bring it to more countries outside of Europe soon. But don't expect it to launch in the EU anytime soon – stricter regulations in the region, like the Digital Markets Act (which already delayed Apple Intelligence there) and the new EU AI Act, are likely keeping it off the table for now. This new feature builds on Microsoft's Copilot assistant by letting the AI actually see and interact with what is on your screen. Basically, you can share an app or your screen and Copilot can then help you with stuff like learning a new app, giving real-time tips while gaming or even analyzing a webpage or photo you are looking at. When you turn it on, Copilot Vision basically becomes your second set of eyes. It watches what you are doing and can chime in with answers, insights, or guidance – kind of like a smart buddy who's always there when you need help staying focused or figuring something out. Whether you are working, browsing or knee-deep in a project, it is built to keep things flowing smoothly. Video credit – Microsoft The best part? It is completely free (at least for now). You won't need a Copilot Pro subscription to try it out, though it is still part of Microsoft's experimental Copilot Labs projects. You will find it inside the Copilot app and you activate it using a little glasses icon – from there, you can choose which app or browser window to share, similar to how screen sharing works in a Microsoft Teams call. It's worth noting that this is an opt-in feature. You have to specifically allow Copilot to see your apps – it's nothing like Microsoft Recall, which automatically takes snapshots in the background once it is turned on. This update pushes Copilot closer to becoming a full-on digital sidekick. I mean, Microsoft clearly wants it to be your go-to assistant for everyday tasks. That said, it is fair to raise a few eyebrows about security. Even though Copilot Vision only activates when you allow it and doesn't interact with anything unless you give it the go-ahead, there are still concerns – especially for anyone working with sensitive data. Personally, I'd skip using it for anything private or confidential. Games and maybe some photo editing? Sure. But beyond that? Not really worth the risk, especially considering how attractive Microsoft is as a target for hackers. Copilot Vision is available on both Windows 10 and Windows 11, but let's not forget Microsoft's been pushing hard for people to switch to the newer OS. So, while it is technically supported on Windows 10 for now, that probably won't last much longer – just like the OS itself.
