Yazılar

Meta to Cut 600 Jobs in Superintelligence Labs as AI Unit Restructures

/in /tarafından

Meta announced plans to cut approximately 600 positions within its Superintelligence Labs division, part of a restructuring aimed at making the company’s artificial intelligence operations more agile and efficient. The layoffs will impact teams across Facebook Artificial Intelligence Research (FAIR), product-related AI, and AI infrastructure, according to the company.

Meta said affected employees are encouraged to apply for other internal roles. However, the newly created TBD Lab — a smaller group of researchers and engineers developing next-generation foundation models — will remain untouched. Chief AI Officer Alexandr Wang emphasized that the reduction in staff would streamline decision-making and increase each member’s scope and influence.

The reorganization follows a period of leadership turnover and mixed reception to Meta’s Llama 4 open-source model. The company recently consolidated all AI initiatives under the Superintelligence Labs umbrella to accelerate progress in foundational and applied AI research.

Separately, Meta secured a $27 billion financing agreement with Blue Owl Capital to fund its largest data center project to date. Analysts say the deal could help Meta advance its massive AI infrastructure plans while mitigating financial risks.

India Proposes Tough AI Labelling Rules to Curb Deepfakes and Misinformation

/in /tarafından

India’s government has unveiled draft regulations requiring artificial intelligence and social media platforms to clearly label AI-generated content, in a sweeping effort to combat deepfakes and misinformation amid rising concerns over the technology’s misuse.

The proposed rules, released Wednesday by the Ministry of Electronics and Information Technology, would compel companies such as OpenAI, Google, Meta, and X to include visible AI markers covering at least 10% of a video or image’s surface area, or the first 10% of an audio clip’s duration, to indicate that the material was artificially created.

India — home to nearly 1 billion internet users — has faced an explosion of AI-generated deepfakes and false information, particularly during elections, in a country already divided along ethnic and religious lines. Officials warn that manipulated videos and fake news could incite violence and erode public trust.

Under the proposal, platforms must also ask users to declare whether their uploads are AI-generated and introduce technical safeguards to verify authenticity. The ministry said the rules aim to ensure “visible labelling, metadata traceability, and transparency for all public-facing AI media.”

The government cited a growing threat from generative AI tools capable of impersonating individuals, spreading propaganda, or manipulating elections. “The potential for harm has grown significantly,” it said in a statement inviting public and industry feedback by November 6.

Legal experts noted that the new labelling rule is one of the first in the world to set a quantifiable visibility standard. Dhruv Garg, founding partner of the Indian Governance and Policy Project, said it would require AI platforms to develop automated detection and tagging systems that identify synthetic content at the moment of creation.

The issue has already reached India’s courts. Bollywood actors Abhishek Bachchan and Aishwarya Rai Bachchan recently sued to block AI-generated videos using their likenesses, while challenging YouTube’s AI training policies.

India’s fast-growing digital landscape has made it a major market for AI firms. OpenAI CEO Sam Altman said in February that the country is the company’s second-largest market by user numbers, which have tripled in the past year.

UN Cybercrime Pact to Be Signed in Hanoi Sparks Both Hope and Human Rights Concerns

/in /tarafından

A landmark United Nations cybercrime treaty, designed to strengthen global cooperation against online offences costing the world economy trillions of dollars each year, is set to be signed this weekend in Hanoi, marking a major step in international cybersecurity governance — but also igniting deep concerns about human rights risks.

The UN convention, which will take effect once 40 nations ratify it, aims to accelerate cross-border responses to crimes such as ransomware, phishing, and online trafficking. However, human rights groups, major technology firms, and even the UN High Commissioner for Human Rights have warned that vague definitions in the treaty could allow authoritarian governments to misuse the pact for surveillance or censorship.

The European Union and Canada have confirmed plans to sign, saying the final text includes safeguards for civil liberties, while the U.S. has not confirmed whether it will attend the signing ceremony. UN Secretary-General António Guterres is scheduled to preside over the event on Saturday.

Vietnam’s role as host has drawn scrutiny due to its record of online repression. The U.S. State Department and Human Rights Watch recently reported that at least 40 people have been arrested in Vietnam this year for online posts critical of the government. Critics say holding the signing there “sends a troubling message” about digital rights, particularly as Vietnam continues to tighten control over internet speech.

The Cybersecurity Tech Accord, a coalition that includes Meta and Microsoft, has dubbed the agreement a “surveillance treaty,” warning it could enable excessive data sharing between governments and “make it easier, not harder, for criminals to engage in cybercrime.”

Despite the controversy, the UN Office on Drugs and Crime (UNODC), which led negotiations, insists the treaty includes human rights protections and allows countries to refuse cooperation requests that violate international law. It also states that the agreement “encourages legitimate cybersecurity research” — a point activists fear could still be used against ethical hackers who expose government vulnerabilities.

Vietnamese officials defended hosting the event, saying the nation faces rising cyberattacks on critical infrastructure and hopes the accord will boost its cyberdefence capabilities. Still, digital rights advocates like Raman Jit Singh Chima of Access Now warn that the pact risks being “a tool for repression disguised as global cooperation.”