EU Tells Instagram, Facebook to Change Addictive Features or Risk Fineshttps://english.aawsat.com/technology/5294432-eu-tells-instagram-facebook-change-addictive-features-or-risk-fines
EU Tells Instagram, Facebook to Change Addictive Features or Risk Fines
FILE PHOTO: A blue verification badge and the logos of Facebook and Instagram are seen in this picture illustration taken January 19, 2023. REUTERS/Dado Ruvic/Illustration/File Photo
EU Tells Instagram, Facebook to Change Addictive Features or Risk Fines
FILE PHOTO: A blue verification badge and the logos of Facebook and Instagram are seen in this picture illustration taken January 19, 2023. REUTERS/Dado Ruvic/Illustration/File Photo
The EU charged Meta Platforms' Instagram and Facebook on Friday with breaching its tech rules, with regulators targeting features they say are designed to keep users hooked and demanding changes to autoplay and infinite scroll or risk fines.
The European Commission's preliminary findings follow a two-year investigation under the European Union's landmark Digital Services Act, which requires large online platforms to do more to tackle illegal and harmful content.
Social media companies face growing scrutiny around the world over concerns that their platforms are contributing to a mental health crisis among children, prompting some governments to impose or consider bans for underage users.
The Commission, the EU's tech regulator, said Meta had failed to adequately assess the addictive risks posed by highly personalized recommendations, autoplay and infinite scroll, which continuously feed users new content and encourage prolonged engagement.
It said reels and stories on Facebook and Instagram could contribute to excessive or compulsive use.
The regulator criticized Meta's measures to mitigate these risks, saying time management tools can be easily dismissed, while parental controls require significant time, effort and technical knowledge to use effectively.
Meta should disable features such as autoplay and infinite scroll by default, introduce effective screen-time breaks and make its recommendation system less focused on driving engagement, Reuters quoted the Commission as saying.
"We disagree with these preliminary findings, which don't accurately take into account the significant steps we've taken to protect teens," Meta spokesperson Ben Walters said.
"Since this investigation began, we rolled out Teen Accounts that automatically protect teens and put parents in control - allowing them to block access to Instagram at night and cap daily screen time at just 15 minutes."
Meta added it would continue to engage constructively with EU regulators.
"Our starting point is that, based on our findings, this design is too addictive and changes need to be made," EU tech chief Henna Virkkunen told Reuters.
"The next step is either that Meta changes its design or a non-compliance decision will follow."
Meta, which risks a fine of up to 6% of its global annual turnover, can respond to the charges before the Commission issues a final decision in the coming months.
The company last month failed in its bid to dismiss claims by 29 US state attorneys general's that Facebook and Instagram are addictive to children.
The EU charges against Meta mirror those brought against TikTok in February, when regulators demanded similar changes to its app.
The Commission is separately investigating so-called rabbit hole effects caused by Facebook and Instagram recommendation systems, where users can be drawn into prolonged viewing by algorithmic recommendations that push them towards similar content. In another case announced in April, it told Meta to do more to prevent children under 13 from accessing its social networks or risk fines.
The Commission is due to receive findings from experts on Monday that could help pave the way for a Europe-wide social media ban for teenagers that Commission President Ursula von der Leyen is expected to announce in her September state of the union address.
Global AI Industry Falls Short on Safety, Think Tank Warnshttps://english.aawsat.com/technology/5294420-global-ai-industry-falls-short-safety-think-tank-warns
Global AI Industry Falls Short on Safety, Think Tank Warns
Mistral ranked last in a survey on the management of risks associated with artificial intelligence. Lionel BONAVENTURE / AFP/File
US artificial intelligence lab Anthropic scored the highest in a semiannual safety ranking, but globally the industry fails to combat "existential" threats, according to a report released on Tuesday.
Meta moved up two spots to fourth place, while xAI dropped three spots to seventh place, just ahead of China's DeepSeek and France's Mistral, which placed last, according to US-based AI safety think tank Future of Life Institute, which ranked nine of the world's leading AI companies.
Seven researchers and governance experts determined the rankings based on public data and information provided by the companies.
They evaluated efforts across six distinct categories: risk assessment, current harms, safety frameworks, existential safety, governance and accountability, and information sharing.
No company received an "A" in any single category, while Anthropic got the best overall score of "C+."
Mistral was included on the list for the first time, though when asked by AFP to comment on its last place, the company said the report's framework isn't suited for its approach to developing AI models.
The French company develops so-called open models, which allow users to download and modify them. Many of its competitors develop closed AI models -- including Anthropic, OpenAI and Google DeepMind, which are also included in the report.
"I was very disappointed to find that they came last, especially since Europe has really...been a leader in AI safety," Max Tegmark, an MIT professor and Future of Life president, told AFP.
"We reached out many, many times" but Mistral did not respond to the organization's survey, Tegmark continued.
Alibaba, xAI and DeepSeek did not respond to its survey either, the organization said.
Three Chinese developers included in the report also produce open models and landed in the bottom half of the ranking: DeepSeek (fifth), Alibaba Cloud (sixth) and Z.ai (eighth).
- 'Questionable' practices -
The report noted that several companies that previously banned their technology from military uses have "gradually reversed course," including Anthropic, which the report criticized for having "questionable military engagements."
The US government used Anthropic's technology in military operations in Venezuela and Iran over the past year, according to various media reports -- though the company was subject to a recent ban by the Pentagon over disagreements on AI safety.
All nine companies are failing when it comes to combating "existential" threats such as pursuing models that reach human-level intelligence, known as "artificial general intelligence" or AGI, the report said.
Although "constructive attempts exist," efforts across the board are "entirely inadequate."
Other risks include the possible misuse of a model to carry out a cyberattack or perform tasks potentially harmful to humans.
Anthropic was thrust into the spotlight recently after it released its most powerful model yet, called Mythos.
In early April, the San Francisco-based company released Mythos only to a handful of trusted organizations due to its abilities to expose cyber safety vulnerabilities to bad actors.
However, by June 12 the US government blocked Anthropic from releasing Mythos to foreigners on national security grounds.
The Trump administration eventually lifted the ban a couple of weeks later on June 30.
Humanitarians Look to Put the AI in Aidhttps://english.aawsat.com/technology/5294364-humanitarians-look-put-ai-aid
The World Food Program is using autonomous trucks to deliver aid in South Sudan, Sudan and Uganda. Rian COPE / AFP
From remote-controlled trucks delivering life-saving aid in dangerous settings to mobile phone data analysis flagging mass displacement, humanitarians are eyeing ways in which artificial intelligence can speed up and improve their operations.
There have been plenty of warnings about the dangers of AI for aid agencies, who face growing challenges of securing often extremely sensitive data and swelling misinformation about their operations and beneficiaries, said AFP.
But at the AI for Good summit in Geneva this week, a handful of humanitarian-focused displays emphasized the technology's positive potential.
Parked in one corner of a vast hall at the Palexpo conference center was a giant white SHERP vehicle, resembling a hulking Martian rover, decked out with cameras and sensors and a drone landing-pad on the roof.
Made in Ukraine, SHERPs are amphibious vehicles that can float on water, drive through swamps and flooded rivers with their giant wheels, and climb over obstacles up to one meter (3.3 feet) high.
The UN's World Food Program is preparing to begin field-testing a version of the AI-enabled truck that can be steered remotely through the most dangerous and difficult terrain to reach people in need.
"I think this could be a game-changer," Bernhard Kowatsch, head of WFP's global accelerator and ventures innovation division, told AFP.
The technology, he said, "should allow us essentially to reach people that otherwise never would have been reachable".
Not possible without AI
WFP already has drivers using SHERPs to deliver aid in Sudan, South Sudan and Uganda.
But after numerous heartbreaking losses of drivers, it tasked the German Aerospace Center (DLR) to help equip the vehicles with AI and other technologies, making it possible to control them remotely through particularly dangerous terrain.
The idea is to set up a shipping container control room in a safe area, where a human can remotely control the vehicle on the last, most treacherous leg of its journey.
Tests have been conducted in Germany, and will be tried out in the field in Uganda in 2028, said Armin Wedler, who is coordinating DLR's Autonomous Humanitarian Emergency Aid Devices (AHEAD) project.
Standing next to the 2.8-meter high vehicle, he told AFP that the team had used "remote-control technologies which are based on mathematics and old-school... research", but stressed: "We would not be able to process everything without using also AI".
It would be possible to make the vehicle fully autonomous, Wedler said, but stressed that in complex humanitarian settings "we have to have a human in the loop".
"We're not talking about driving on clear streets with clear lanes. There are no streets," he said, also describing scenes where aid trucks are suddenly swarmed by desperately hungry people.
"There's no AI autonomous algorithms ever capable to handle that safely."
'Life-saving'
Among more than 200 exhibitors at the summit -- showing off everything from humanoid robots to bionic prosthetics and emotional companions -- the other humanitarian displays were more discreet, with pamphlets detailing how AI tools are boosting and streamlining operations.
Among them, the UN refugee agency detailed a new Legal Virtual AI Assistant for lawyers and legal officers representing refugees, enabling them to swiftly determine the rights available within country-specific legal frameworks.
Rebeca Moreno Jimenez, the lead data scientist at UNHCR's Innovation Service, told AFP that building cases faster and more efficiently can be "life-saving for many refugees".
Another UN initiative called Data Insights for Social and Humanitarian Action, or DISHA, relies on partnerships with private actors such as Google and McKinsey to provide humanitarian organizations with data and AI models to speed up and improve disaster responses.
One project uses AI analysis of anonymized mobile phone data to spot mass-population movements during disasters, determining where people are fleeing, to help humanitarians better tailor their response.
Another uses AI for rapid analysis of satellite images taken before and after disasters like last month's earthquakes in Venezuela to determine building damage.
The aim is to give humanitarians "accurate information early enough to make better decisions (and) avoid going to the wrong place when there are people who need you somewhere else", DISHA product lead Andreas Kortis told AFP.
How AI Supports Real-Time Decision-Making in Saudi Arabia's Airports and Portshttps://english.aawsat.com/technology/5294121-how-ai-supports-real-time-decision-making-saudi-arabias-airports-and-ports
Airports and ports require an intelligent operational layer that connects data, processes, and resources to support better decision-making during disruptions. (Adobe)
How AI Supports Real-Time Decision-Making in Saudi Arabia's Airports and Ports
Airports and ports require an intelligent operational layer that connects data, processes, and resources to support better decision-making during disruptions. (Adobe)
Infrastructure investment in Saudi Arabia is entering a new phase, one measured not only by the airports, ports, logistics corridors, energy systems, and digital infrastructure that have been built, but also by the ability of these assets to operate as a single integrated system.
The next source of value will not come solely from expanding capacity, but from improving the decisions that determine how aircraft, ships, cargo, energy, and data move from one moment to the next. At this stage, artificial intelligence becomes part of the operational equation itself: how can major infrastructure assets be transformed into operations that are more efficient, reliable, and resilient in the face of disruptions?
Bilal Abu-Ghazaleh, Founder and CEO of 1001. (Company)
In an exclusive interview with Asharq Al-Awsat, Bilal Abu-Ghazaleh, founder and CEO of 1001, a startup developing sovereign artificial intelligence that recently raised $30 million in a Series A funding round, said the difference between building assets and operating them efficiently is the difference between capacity and performance.
"Building an airport, a port, or a logistics corridor gives you capacity, but it does not automatically ensure the best use of it," he said. "A larger asset does not manage itself more efficiently. Instead, it creates a greater number of decisions that must be made correctly."
Intelligence Built into the Design
This idea lies at the heart of understanding the next phase of Saudi Arabia's transformation. Every new asset adds not only physical space or operational capacity, but also an entirely new network of relationships and interdependencies. A new airport terminal, an additional port berth, or a new logistics corridor does not function in isolation from the rest of the system. A delay involving a single aircraft or vessel can alter gate or berth assignments, triggering a chain reaction that affects truck movements, customs operations, warehouses, delivery schedules, and workforce allocation.
Abu-Ghazaleh explained that adding a new terminal, berth, or corridor also means "adding thousands of new connections between things that influence one another." Given the speed and scale of Saudi Arabia's development, he said, no human team, regardless of its experience, can keep track of all these relationships and consistently make the best decisions in real time.
Yet this challenge also presents an opportunity. Countries with aging infrastructure are often forced to introduce artificial intelligence after decades-old systems are already in place. Saudi Arabia, by contrast, can embed an intelligent operational layer into new projects such as King Salman International Airport and newly developed ports and rail networks from the design stage rather than years after operations begin.
Abu-Ghazaleh argued that "most countries around the world are stuck trying to bolt artificial intelligence onto legacy systems," whereas Saudi Arabia has the opportunity to design intelligence directly into its infrastructure assets from the outset.
Solving Complex Problems
At airports and ports, the most difficult challenges are not always a lack of capacity but a lack of coordination. That is why building more facilities, hiring more staff, or introducing another conventional software system is not enough. When a vessel is delayed at a major port such as Jeddah Islamic Port, or when a disruption occurs at a large airport, it sets off a chain of decisions: Which berth should be assigned? How should cranes be rescheduled? What happens to trucks and rail operations? How should yards, warehouses, and resources be reorganized?
Abu-Ghazaleh said, "The hardest problems in airports and ports are not capacity problems. They are coordination problems." He stressed that these cannot be solved simply by pouring more concrete or increasing the workforce. Adding more people can actually increase the coordination burden without necessarily providing a unified view of all the interconnected variables.
Conventional software also cannot fully bridge the gap because the core issue lies in fragmented data spread across multiple systems: one for transportation, another for warehouses, a third for enterprise resource planning, and others for customs or maintenance. Each system performs a specific function within its own domain, but none has visibility across the entire operation. The challenge, therefore, is to build a living operational model that unifies data, relationships, and business rules, enabling decisions to be made based on a single, comprehensive view of the system.
Deploying artificial intelligence in critical infrastructure requires human oversight, explainability, auditability, and a record of every decision. (Shutterstock)
The Role of Operational Intelligence
Abu-Ghazaleh emphasized that there is no single starting point for every industry. The greatest benefit from artificial intelligence may lie in improving capacity utilization at an airline, managing disruptions at a port, optimizing cargo flows across a logistics company, or reducing energy consumption at another infrastructure asset. "We don't start by guessing," he said. "We start by understanding the operation from the inside."
According to Abu-Ghazaleh, 1001's methodology places engineers within clients' teams to understand how an organization actually operates, rather than how it appears in diagrams or presentations. These engineers map workflows, data flows, and the highest-value operational challenges before working with operations teams to identify the first use case capable of delivering a measurable impact. They then build what he describes as a "living operational model," essentially a dynamic digital map that captures assets, processes, business rules, and the relationships between them, while updating continuously in real time.
The significance of this approach is that the value extends far beyond a single use case. Once this foundation has been established, subsequent applications can be developed much more quickly. Abu-Ghazaleh noted that while the first use case typically takes the longest to complete, the second and third benefit from the same underlying model, reducing implementation time from around 16 weeks to roughly four weeks. He added that the returns can be substantial, with a single use case capable of generating more than $100 million in value during its first year.
When a Ship Is Delayed
To illustrate the difference between automation and intelligence, Abu-Ghazaleh uses the example of a ship arriving several hours behind schedule. Such an event does not simply alter one arrival time. It disrupts the entire operational plan. The berth assigned to that vessel may now be needed for another ship, while the cranes and crews waiting for it remain idle. The containers it carries are linked to trucks, trains, and delivery schedules that are no longer aligned, even as the yard has already been organized according to the original arrival timetable.
Automation can handle some routine tasks, such as sending an alert, updating a schedule, or reallocating a slot based on predefined rules. But when real-world conditions diverge from the plan, executing a fixed rule is no longer enough. What is required is a complete reassessment of the operation and the identification of the best recovery plan across thousands of variables within minutes.
Abu-Ghazaleh said this is precisely the type of decision that artificial intelligence can improve because it "sees the entire operation at once." It can rapidly develop a new operating plan by determining which berth should accommodate the delayed vessel, how cranes and yard operations should be reorganized, and how trucks and trains should be rescheduled together while accounting for real-world constraints. Even so, this does not eliminate the role of human operators. The proposed plan is presented to the operator along with the reasoning behind it, while the final decision remains under human control.
AI as Part of the Operation
The risks change when artificial intelligence moves beyond analysis into recommendation or execution. At the analysis stage, the system serves as an advisory tool. If it makes a mistake, a human can identify the error before any harm occurs. But once it begins making recommendations that influence the operation of an airport, port, or energy asset, it becomes part of the operational process itself.
Abu-Ghazaleh explained that "failure in these environments is not a software bug in a report. It is a crisis." For that reason, the standard for trust becomes significantly higher. He identified three essential requirements: consistent reliability, the ability to explain the reasoning behind every recommendation or action, and the recording of every step so that decisions can be audited and reversed when necessary.
He added that the system must be governed rather than trusted unconditionally, while human operators must remain in control. The system should earn greater autonomy gradually, "one decision at a time."
This also explains why many AI projects struggle to move from the pilot stage to live operations. The gap is not merely technological. It is also operational and institutional. There is a significant difference between a convincing demonstration and a system that an operator is willing to trust alongside an airport runway or on a port berth. As a result, system governance, auditability, and operational continuity become essential requirements for deployment in critical infrastructure rather than optional features that can be added later.
The Challenge of Incomplete Data
Large-scale operations often contend with fragmented or incomplete data, particularly when they rely on legacy systems. Abu-Ghazaleh believes the answer is not to wait for perfect data, but to build integration and data lineage instead.
"You will never get perfectly clean data from systems that are decades old, and you do not need to," he said. The objective is to bring together scattered data sources within a consistent operational model while ensuring that every recommendation can be traced back to the data, logic, and alternative options on which it was based.
He added that data alone cannot capture everything that happens within an operation. A significant portion of operational knowledge exists only in the minds of the people who manage these systems every day and is not stored in any database. According to Abu-Ghazaleh, 1001 uses AI agents to capture that context directly from operational teams and integrate it into the model alongside system data. When data gaps or conflicts between sources emerge, they should be made visible to the operator rather than concealed, and the decision should be escalated to a human instead of having the system rely on guesswork.
Sovereignty Alone Is Not Enough
Sovereignty remains an important part of the discussion, particularly when it comes to critical infrastructure. However, Abu-Ghazaleh cautions against equating resilience with local deployment alone, arguing that "deploying systems locally does not automatically make them resilient." In his view, two distinct risks must be addressed: the risk of foreign control, and the risk of depending on a single component, facility, or supplier, even within the domestic market.
From this perspective, sovereignty addresses the risk of an external "kill switch" by ensuring that data, models, and infrastructure remain under national jurisdiction. Resilience, however, also requires an architecture that is not locked into a single model, vendor, or location. Abu-Ghazaleh told Asharq Al-Awsat that this is why model- and vendor-neutral infrastructure is essential, along with the ability to operate across both cloud and on-premises environments. He also stressed the importance of dividing systems into distinct components so that the failure of one does not bring down the entire operation.
He summarized the relationship succinctly: "Local control without redundancy and alternatives is fragile, while redundancy without control leaves you exposed." In other words, intelligent national operations require both sovereignty that reduces external dependence and resilient system design that minimizes domestic points of failure.
The Skills That Matter After Computing Infrastructure Is Built
In recent years, many countries have focused on building data centers and expanding computing capacity. Abu-Ghazaleh argues, however, that leadership in artificial intelligence will not be determined by computing power alone, but by applied AI, meaning the integration of intelligence into real-world operations.
"Applied AI is not won in the laboratory," he said. "It is won inside the most demanding live operations."
In his view, Saudi Arabia holds several advantages, including major infrastructure across aviation, ports, energy, and logistics, the ability to move quickly when institutions are aligned, new projects into which intelligence can be embedded from the outset, and the capacity to invest at scale. Turning those advantages into measurable performance, however, requires three capabilities: a data foundation that transforms fragmented systems into a unified decision-making asset, engineers who understand both software and operational environments, and institutional leadership willing to redesign decision-making processes while maintaining governance, security, and auditability.
If these capabilities are in place, computing infrastructure becomes a productive investment. Without them, it risks becoming expensive but underutilized capacity. As Abu-Ghazaleh put it, shortcomings in these areas turn investment into "costly idle capacity."
The Next Measures of Success
Abu-Ghazaleh believes Saudi Arabia's success will be measured by its ability to move beyond building AI infrastructure and begin embedding artificial intelligence into national operations. The true indicators, he said, will be operational rather than experimental.
The first indicator is that AI becomes part of the daily operations of ports, airports, and energy infrastructure instead of remaining confined to pilot projects. "Ninety-five percent of AI never reaches production," he said. The real test is whether day-to-day deployment is reflected in measurable outcomes such as productivity, on-time performance, uptime, and cost.
The second indicator is sovereignty across data, computing infrastructure, and AI models. When critical systems operate using intelligence built on the Kingdom's own data and models, that intelligence becomes a strategic asset that grows with every decision rather than a service that can be priced, restricted, or switched off from abroad.
The third indicator is ensuring that expertise remains within the domestic market through Saudi engineers and accumulated applied experience gained from each successful deployment.
In this sense, the next phase is not simply about adding artificial intelligence to existing infrastructure. It is a test of whether the entire system can transform major physical assets into smarter, more intelligent operations.
لم تشترك بعد
انشئ حساباً خاصاً بك لتحصل على أخبار مخصصة لك ولتتمتع بخاصية حفظ المقالات وتتلقى نشراتنا البريدية المتنوعة
