A coalition of major publishers — Elsevier, Cengage, Hachette, Macmillan, and McGraw-Hill — along with bestselling author Scott Turow and his company S.C.R.I.B.E., Inc., filed a class action complaint against Meta Platforms, Inc. and its CEO, Mark Zuckerberg, in the U.S. District Court for the Southern District of New York on May 5, 2026.

The plaintiffs allege that the defendants committed extensive copyright infringement by downloading pirated books and journal articles, scraping unauthorised web content and repeatedly using these materials to train successive versions of the Llama AI model, all without paying licensing fees.

According to the complaint, the allegedly infringed material spans textbooks, scientific journals, and popular fiction, including titles such as The Fifth Season by N. K. Jemisin and The Wild Robot by Peter Brown.

Meta Defends AI Training as Fair Use: A Meta spokesperson dismissed the suit, arguing that courts have already found AI training on copyrighted material can qualify as fair use, and vowed to fight the lawsuit aggressively.

The company’s defence is part of a broader legal strategy that has recently succeeded in U.S. courts. In June 2025, a U.S. federal court ruled in Meta’s favour in a separate lawsuit brought by authors who claimed the company unlawfully used copyrighted books to train its Llama AI models. The court found that the plaintiffs did not sufficiently prove market harm from Meta’s AI systems, a key factor in determining fair use under U.S. copyright law.

The judge ruled that Meta’s use of the books was “highly transformative,” and noted that Llama did not reproduce substantial portions of the original works, even with adversarial prompting. The court also stated that although Meta’s AI business is commercial, commercial intent alone does not preclude a fair use defense. The court also addressed Meta’s alleged use of pirated books separately from the fair use analysis. The judge stated that downloading from shadow libraries does not automatically invalidate a fair use defense, but noted that using pirated repositories may indicate bad faith in certain situations.

Allegations Around Meta’s Use of Pirated Books: In November 2025, Entrepreneur Media sued Meta, alleging the company used pirated articles to train its Llama AI model. The complaint, filed in the U.S. District Court for the Northern District of California, claimed that several Entrepreneur titles and registered magazine issues, including Start Your Own Import/Export Business, appeared in the LibGen corpus. Meta allegedly downloaded this corpus with Mark Zuckerberg’s authorization to train Llama. Entrepreneur argued that the harm was direct and commercial, as users could prompt Llama to generate similar guidance for free, reducing demand for its business guides and periodicals.

In June 2025, a U.S. federal judge granted Meta summary judgement in the Kadrey v. Meta case, dismissing copyright infringement claims related to AI training due to insufficient evidence of market harm. However, the case continued on separate claims regarding the alleged distribution of books obtained through torrent downloads. This ruling, which publishers are watching closely, follows Anthropic’s $1.5 billion settlement with authors in September 2025, highlighting the significant financial risks for large language model developers. The varied outcomes in these cases make the new publishers’ class action a critical test of whether courts will hold AI companies liable for the entire process, from acquiring pirated content to producing outputs that may copy-protected text.

Six Legal Counts, a Class Action, and Demands for Injunctions: The complaint brings six counts against Meta and Zuckerberg, covering copyright infringement through torrenting, web scraping, and AI training; distribution of infringing works; contributory infringement by Zuckerberg personally; and violations of the Digital Millennium Copyright Act for stripping copyright management information.

The proposed class covers all legal or beneficial owners of registered copyrights in books with an ISBN or journal articles with a DOI or ISSN that Meta reproduced without authorisation by torrenting or web scraping, distributed by torrenting, or reproduced in connection with training a Llama model.

Plaintiffs seek permanent injunctions to prevent ongoing infringement, statutory damages under the Copyright Act, attorneys’ fees, and a jury trial.

Read more:

• Meta Wins AI Copyright Lawsuit Filed By Authors
• Meta Argues BitTorrent Seeding Is Fair Use in AI Training
• Meta Faces Lawsuit After AI Glasses Recorded Sensitive User Content

The post Major publishers like Macmillan allege Meta used pirated books to train its Llama model appeared first on MEDIANAMA.

The suit, filed on April 3 and reviewed by Reuters, lists 12 Instagram Reels in which Nykaa allegedly used Zee’s licensed music as background tracks to reach its millions of followers.

What is Zee claiming? Zee’s legal argument fundamentally rests on a crucial distinction: its licensing agreement with Meta Platforms permits individuals to use its music in Instagram posts for non-commercial purposes. Nykaa’s use of music to promote products to a large commercial following falls outside that permission, Zee contends.

During a court hearing, Nykaa’s counsel informed the Delhi High Court that all 12 flagged Instagram Reels had been taken down. However, it does not necessarily nullify Zee’s damages claim. None of the companies have commented publicly on the litigation. The matter is next listed for hearing on May 26.

The case highlights tensions that have been building quietly across India’s booming short-video economy. Hindi film songs have become the default currency of brand storytelling on Instagram and YouTube Shorts, with marketing teams routinely selecting tracks from in-app music libraries without scrutinising the commercial use restrictions buried in the platforms’ terms of service.

Aditya Gupta, a partner at Ira Law, said, “Marketing departments often use content available on music libraries without reading the fine print”. The Delhi High Court’s final ruling may force brands to fundamentally reconsider this practice.”

Why it matters: The lawsuit can set an example on how e-commerce platforms engage with licensed content on social media. The availability of a song within a platform’s music library does not, by itself, confer the right to use it in advertising or product promotion. The Meta-Zee music licensing deal explicitly distinguishes between personal and commercial use. Nykaa is alleged to have crossed that line on at least 12 occasions.

More importantly, the outcome of the suit will test whether India’s copyright framework can address the nuances of platform-intermediated commercial content in the age of short-form video marketing.

Zee’s record in IP litigation and comparable cases: Historically, Zee has been aggressive in pursuing copyright infringement claims. In January this year, MediaNama reported that ShareChat had allegedly used Zee Entertainment’s copyrighted material without permission.

In 2024, Zee filed a defamation lawsuit against Bloomberg. Consequently, a Delhi court directed the news publisher to take down an article about Zee’s financials.

Similar legal battles have played out in jurisdictions outside India. In the US, the music industry’s sustained litigation against TikTok’s predecessor Musical.ly, and later against brands using unlicensed music on TikTok itself, forced the platform into sweeping licensing agreements with Universal Music Group, Sony Music, and Warner Music. The UMG–TikTok standoff in early 2024, which saw Universal pull its entire catalogue from the platform over royalty disputes, demonstrated how high the stakes of social-media music licensing had become globally.

India has not yet produced a definitive ruling on the interaction of platform licensing, brand promotion, and music copyright. The outcome will set a precedent for India’s creator and brand economy.

Also Read:

• Zee Accuses ShareChat of Copyright Breach After Licence Expiry; Delhi HC Allows Copyright Case in Delhi
• Delhi Court Directs Bloomberg to Take Down Alleged Defamatory Article on Zee Entertainment Enterprises Ltd
• ZEEL Secures ‘Dynamic+’ John Doe Order to Combat Digital Piracy, Targeting 103 Rogue Websites

The post Zee drags Nykaa to Delhi High Court over copyrighted songs used in Instagram Reels appeared first on MEDIANAMA.

Meta Platforms has introduced new AI-powered age-assurance measures to identify teenagers and underage users on Instagram and Facebook. The company will expand systems that automatically assign suspected teenage users to “Teen Account” protections, even if they registered with an adult birthdate.

Expanded safeguards will be introduced for Instagram users in the European Union and Brazil, and for Facebook users in the United States. The company stated that similar Facebook protections will be available in the UK and the EU in June.

Meta also announced it is strengthening enforcement against users under 13, who are not permitted on its platforms, by using AI systems to analyze behavioral and profile-based signals to identify likely underage accounts.

AI systems will analyze profiles and visual cues: The company stated that its AI systems estimate age using signals such as account activity, profile information, social interactions, and visual indicators from photos and videos. The blog post also suggests that the systems may consider cues such as height and bone structure. Meta clarified that the technology does not use facial recognition to identify individuals.

Accounts identified as belonging to teenagers will automatically receive stricter protections, including restrictions on messaging from strangers, reduced exposure to sensitive content, and limits on certain platform features.

Meta is streamlining underage account reporting by simplifying in-app and Help Center reporting flows. The company is also enhancing its response to these reports. Human review teams are now supported by AI models that apply consistent evaluation criteria to each report. Testing has shown that AI-driven reviews provide greater accuracy and faster resolutions than human reviews alone.

Meta calls for app stores to step up at the industry level: While investing in its own systems, Meta is also advocating for an industry-wide solution. The company believes legislation should require app stores to verify users’ ages and share this information with apps and developers to enable age-appropriate experiences. Meta notes that this approach is already popular with the public, with 88% of US parents supporting it.

Requiring parental approval and age verification at the app store or operating system level, Meta argues, provides a centralised, consistent, and privacy-preserving place for age assurance, rather than requiring every individual app to comply with different rules. The underlying position is clear: no single company can solve the age assurance problem alone.

Why this matters: Meta’s broader use of AI-based age assurance aligns with global efforts to regulate children’s access to social media. Governments across regions are urging technology companies to implement stronger age verification and child-safety measures.

The debate has intensified after Australia enacted a social media minimum age law restricting access for users under 16. Critics cite concerns about privacy, surveillance, and the reliability of age estimation, while supporters believe stronger restrictions are needed to protect minors from harmful content and addictive platform features.

The move comes amid regulatory scrutiny: The announcement comes amid growing global scrutiny of online child safety and age verification practices. Regulators in the European Union recently concluded that Meta failed to adequately prevent children under 13 from accessing Facebook and Instagram, potentially exposing the company to significant penalties under the EU’s Digital Services Act.

Meta is involved in a significant child-safety trial in New Mexico following a $375 million jury verdict. The state is requesting that the court require Meta to block users under 13, delete underage accounts and related data, link minors’ accounts to parents, restrict adult-minor messaging, and limit features such as infinite scroll, autoplay, and school-hour notifications. Meta has stated that these demands are technically impractical and warned that it may withdraw its platforms from New Mexico if a workable solution cannot be found.

Also read:

• “No scientific evidence”: 438 scientists call for pause on age-based controls until benefits and risks understood
• Meta to Use AI to Detect and Move Teens to Protected Accounts
• Meta under fire in the US for allegedly allowing underage users to access its platforms

The post Meta expands AI-based age assurance for teens amid rising regulatory pressure appeared first on MEDIANAMA.

Privacy advocacy group noyb has filed a complaint against LinkedIn, accusing the platform of restricting users’ General Data Protection Regulation (GDPR) access rights while charging for the same data through its premium subscription service.

The complaint, filed before the Austrian Data Protection Authority, centres on LinkedIn’s “Who viewed your profile” feature. LinkedIn tracks profile visits and shows users a list of visitors from the past 365 days as part of its paid Premium offering. According to noyb, the same information should also be available free of cost under Article 15 of the EU’s GDPR, which gives users the right to access their personal data.

Questions Over Profile Tracking: NOYB argues that LinkedIn refused to provide this information when a user filed an access request, while continuing to offer it as a paid feature. The group has also questioned whether LinkedIn’s tracking of profile visits itself complies with GDPR rules, saying users are allowed to opt out but are not asked for explicit consent beforehand.

Wider Scrutiny of LinkedIn’s Data Practices: The complaint comes amid wider scrutiny of LinkedIn’s data practices in recent years. In 2024, Ireland’s Data Protection Commission fined LinkedIn €310 million over GDPR violations linked to behavioural analysis and targeted advertising, saying the company lacked a valid legal basis for processing user data for ads.

Separately, LinkedIn also faced a lawsuit in the US in early 2025 over allegations that it used user data for AI training through default opt-in settings. The company had also updated its privacy policies to expand how user data could be shared with Microsoft and affiliate companies for personalised advertising and AI-related features, prompting concerns from privacy advocates over consent and transparency.

noyb’s Argument Against LinkedIn: “Selling data to its own users is a popular practice among companies. In reality, however, people have the right to receive their own data free of charge,” said Martin Baumann, data protection lawyer at noyb. “It is absurd that companies only seem to recognise the importance of data protection when they want to sell data.”

Noyb further argued that LinkedIn cannot cite privacy concerns to deny access requests if it is already willing to share the same data through a paid subscription.

“The protection of the rights and freedoms of others can definitely be a reason for not disclosing shared personal data. However, if a company has sought the relevant consent and is clearly willing to make the same data available for a fee, this argument no longer holds water,” Baumann said.

What noyb Wants: In a LinkedIn post announcing the complaint, noyb said: “fundamental rights should NEVER be a ‘premium feature.’”

The group has asked regulators to direct LinkedIn to fully respond to the access request and impose a fine to prevent similar violations in future.

Read More:

• LinkedIn Set To Share Users’ Data With Microsoft For Personalised Ads And AI Training
• LinkedIn received 10 government requests from India in 2019, didn’t give up data
• LinkedIn Sued for Alleged Misuse of User Data For AI Training

The post LinkedIn accused of GDPR violations for charging users for profile view data appeared first on MEDIANAMA.

The Securities and Exchange Board of India (SEBI) has named Anthropic’s Claude Mythos in a circular dated May 5, ordering every regulated entity in Indian securities markets to immediately overhaul their cybersecurity infrastructure. SEBI is the first Indian financial markets regulator to name a specific AI model in a formal circular; CERT-In had first named Mythos across all sectors on April 26.

The circular covers stock exchanges, depositories, mutual funds, brokers, credit rating agencies, custodians, merchant bankers, and portfolio managers, among others.

SEBI’s concern: Mythos can identify and exploit vulnerabilities “using speed and scale,” threatening data confidentiality, application integrity, and reliability of outputs. Because all market participants are interconnected, one breach can trigger a domino effect across the entire ecosystem.

SEBI has also constituted a task force called cyber-suraksha.ai, comprising representatives from market infrastructure institutions (MIIs), qualified registrars and transfer agents (QRTAs), and other regulated entities, to examine AI-driven cybersecurity risks, share threat intelligence, report cyber incidents on priority, and review third-party vendor security posture.

What the circular requires:

1. Patch immediately. All operating systems and applications must be updated right away. Where patches don’t exist yet, entities must use virtual patching, a temporary protective layer applied over a vulnerability when no official fix is available.
2. Run AI-based vulnerability assessments. Entities must conduct Vulnerability Assessment and Penetration Testing (VAPT), simulated cyberattacks used to find weaknesses before real attackers do, continuously using both conventional and AI-based tools, in line with SEBI’s Cyber Security and Cyber Resilience Framework (CSCRF).
3. Hold vendors accountable. Entities must engage third-party vendors on timely patching. Exchanges and depositories must specifically direct their empaneled application vendors to run comprehensive risk assessments on AI-led vulnerability detection models and implement safeguards, including patching, VAPT, and continuous monitoring.
4. Lock down change management. Every system change, including minor ones, requires full documentation, impact analysis, structured review, rigorous testing, and secure deployment.
5. Secure all APIs. An Application Programming Interface (API) is a connection point that lets different software systems communicate; in financial markets, APIs connect brokers, exchanges, and payment systems. The circular requires:

• Updated inventory of all APIs and apps using them
• Strong authentication on a least-privilege basis (users get access only to what they strictly need)
• Rate limiting and throttling to detect abuse
• Connections only via whitelist

6. Overhaul SOC monitoring. A Security Operations Centre (SOC) monitors an organisation’s systems for threats around the clock. The circular requires:

• Daily monitoring of all systems and networks including low-priority alerts that are typically ignored
• Implementation of a SOAR (Security Orchestration, Automated Response) playbooks integrated with SIEM (Security Incident and Event Management) solutions for automated threat detection and response
• Fast-track onboarding onto the Market SOC (M-SOC), a centralised 24×7 security platform run by NSE and BSE
• MIIs to run workshops to help entities integrate with M-SOC

7. Include AI as a risk scenario. Periodic risk assessments must now explicitly model AI model capabilities as a threat scenario.

8. Harden systems. Entities must adopt secure configurations, disable unnecessary services and default accounts, and enforce Zero Trust Network Architecture (ZTNA), a security model that requires verification at every step, assuming no user or system is trustworthy by default.

9. Update software inventory. Entities must periodically update their Software Bill of Materials (SBOM), a complete list of all software components, including open-source code, for all critical applications.

10. Build long-term AI defence. All regulated entities must prepare a long-term plan for using AI in detection and autonomous agentic mitigation, where AI systems independently identify and respond to threats without waiting for human instruction, including AI-augmented SOC transformation.

Why this matters: SEBI naming Mythos in a circular is a significant regulatory moment. As MediaNama has reported, no Indian company, bank, or government agency has secured access to Mythos under Project Glasswing, Anthropic’s $100 million restricted access programme. MeitY Secretary S. Krishnan confirmed on April 28 that India is still working out logistics with US authorities.

This creates a structural contradiction: SEBI is ordering Indian financial institutions to defend against a model they cannot access to defend themselves with. Claude Security, Anthropic’s enterprise defensive tool, gives Indian firms an indirect path through Infosys as a named partner, but runs on Opus 4.7, which produces two working exploits on the Firefox 147 benchmark against Mythos’s 181, a 90x capability gap.

MediaNama founder Nikhil Pahwa identified the core problem: “A tool that compresses attack timelines without compressing defense timelines increases systemic risk before it improves security.”

The data localisation conflict also remains unresolved. India’s 2018 rules require payment system providers to store all transaction data on servers within India, while Mythos is hosted on US-based servers. The National Payments Corporation of India (NPCI) has not publicly addressed this.

Finance Minister Nirmala Sitharaman chaired a meeting on April 23 with Reserve Bank of India (RBI), NPCI, Indian Banks’ Association (IBA), and CERT-In officials, calling the Mythos threat “unprecedented.” CERT-In has separately issued a high-severity advisory directing organisations to treat every newly disclosed vulnerability as exploitable within hours, not weeks.

Also read:

• India seeks access to Claude Mythos amid cybersecurity fears
• Claude Mythos puts India on alert: CERT-In, telcos, banks assess unprecedented cyber risks
• Anthropic launches Claude Security for enterprises, but stops short of Mythos-level capabilities

The post SEBI forms task force, orders immediate cybersecurity overhaul amid Claude Mythos concerns appeared first on MEDIANAMA.

The Delhi High Court on April 28, TechWiser (founded by Mrinal Saha, with Pratik Rai as on-screen host, 2.52 million subscribers) and TechBar (run by Sanchit Shokeen, 5.72 million subscribers) from making disparaging statements about AI+ Smartphones and its founder Madhav Sheth, on any social media or digital platform, without hearing either channel before passing the order. The order also includes a John Doe clause, which extends the restraint to unnamed future critics, covering speech that has not yet occurred. The next hearing is October 5.

Part of a growing pattern: This is the second ex-parte disparagement injunction, meaning an order passed without hearing the other side, against YouTube critics of a consumer tech brand in as many weeks. In April, Motorola obtained a similar order from a Bengaluru court naming 17 YouTube channels and all major platforms, with creators learning about it only after X’s support team emailed them.

Apar Gupta, founding director of the Internet Freedom Foundation (IFF), told MediaNama at the time that John Doe clauses were designed for piracy cases involving genuinely unidentifiable infringers and their migration into defamation litigation marks a significant expansion of judicial power. “They structurally incentivise over-removal of online criticism, with the user whose speech is taken down having the least effective remedy of any party in the system,” he said.

About AI+ Smartphones: NxtQuantum Shift Technologies operates AI+ Smartphones. Founder Madhav Sheth previously co-founded Realme India in 2018 before departing in June 2023. He launched AI+ in July 2025, marketing it around data sovereignty, the claim that Indian user data stays in India and is not shared with external parties. Per the company, phones run on NxtQuantum OS and are manufactured at United Telelinks’ Noida facility, with data stored on MeitY-approved Google Cloud infrastructure.

What the videos said: TechWiser’s “[This Indian Phone Is A Marketing Disaster!]” posted April 9 had 2.59 lakh views, and TechBar’s “[FAKE Indian Company – Needs to STOP]” posted April 14 had 3.32 lakh views at the time of filing.

Key findings from TechWiser’s hands-on review of the AI+ Pulse 2:

• Pratik Rai connected the phone to a laptop and ran ADB (Android Debug Bridge) commands, a developer tool that accesses parts of a phone’s software not visible to regular users, and found two Chinese apps from Shanghai-based ProCom Technologies, whose terms state it collects user name, profile picture, and phone number. The apps did not appear in the regular app drawer but showed up through ADB. AI+ had previously claimed to have removed these apps after earlier controversies flagged by channel Gyantherapy.
• The privacy dashboard, a feature AI+ markets as showing users which apps access their data, did not display Google app activity. Maps and Gboard accessed location and microphone without appearing on the dashboard.
• AI+ Smartphones’ own privacy policy states: “We may share personal information internally within affiliates, related companies and with other third parties, including Credit Bureaus, for purposes of providing products and services offered by them, such as, personal loans, insurance, deferred payment options, and Pay later through its lending partners.”

TechBar’s findings per its video transcript, as filed before the court:

• TechBar alleged the AI+ Nova Flip shared identical battery, processor, and camera specs with the Chinese Nubia Flip 2.
• TechBar alleged AI+ wearables were identical to products by Chinese company AI Power, whose logo, TechBar said, was nearly identical to AI+.

TechWiser’s video also raised questions about Sheth’s track record. According to TechWiser, he previously led HonorTech India, which stopped launching phones after 2024 and went silent despite denying shutdown rumours, and Nexel India, which holds Alcatel’s rights in India and similarly stopped launching products after a few models. MediaNama has not independently verified these claims.

How the court reasoned: Justice Tushar Rao Gedela reviewed the video transcripts, not the devices, and found the conclusions lacked any formal technical examination by a credible agency. He applied the disparagement test from Dabur India vs Colortek Meghalaya, a legal standard requiring a plaintiff to show that a statement is untrue or misleading, made maliciously, and caused financial damage. At the interim stage the court assesses this only prima facie, meaning it checks whether a case appears strong enough on first look rather than whether it will succeed at trial.

The court found that the statements bordered on disparagement and that the untested analysis had the potential to cause financial loss to AI+ Smartphones. The court reached this conclusion by reviewing the same transcripts, without conducting any independent technical examination of the devices.

The San Nutrition precedent: The order itself cites San Nutrition Pvt Ltd vs Arpit Mangal to lay out the disparagement test. In that 2025 case, the Delhi High Court refused to grant an injunction against influencers who posted critical product reviews, finding that lab reports supported the influencers’ claims and that their defences of truth and fair comment were not clearly untenable. TechWiser’s videos included hands-on ADB testing and direct references to AI+ Smartphones’ own privacy policy.

Safe harbour implications: The suit names YouTube as a defendant alongside the channels. Under the Supreme Court’s ruling in Shreya Singhal vs Union of India, Section 79 of the IT Act protects platforms like YouTube from liability for user-posted content, a protection known as safe harbour. Platforms lose that immunity only upon receipt of a court order or government notification. Gupta told MediaNama in the context of the Motorola case that naming platforms at the outset of a private defamation suit strips safe harbour before any final determination, and that this provides a template for consumer electronics brands to use litigation as a content moderation tool.

MeitY’s draft IT Amendment Rules, 2026 propose making compliance with government advisories mandatory for platforms to retain Section 79 protections.

The John Doe problem: Beyond TechWiser and TechBar, the order restrains unnamed future defendants under the Ashok Kumar clause, a John Doe mechanism originally developed for copyright piracy cases where infringers were genuinely unidentifiable. MediaNama has tracked its use expanding into defamation suits, with such orders leading to collateral takedowns of unrelated content. Any person posting similar criticism of AI+ Smartphones now faces potential legal action without prior notice.

Defendants can challenge the order under Order XXXIX Rule 4 of the CPC (Code of Civil Procedure, 1908), a provision allowing a party to seek vacation or modification of an ex-parte injunction on grounds they were not heard or that the balance of convenience was misapplied.

What AI+ Smartphones said: AI+ Smartphones told MediaNama: “There is a clear distinction between fair criticism and deliberate defamation. The content lacked any credible technical basis or independent verification. We do not seek to silence legitimate reviewers or suppress honest feedback.”

Also Read:

• Motorola obtains ex-parte injunction against 17 YouTube channels in Bengaluru court
• MeitY’s draft IT Amendment Rules 2026: What they mean for platforms, creators and safe harbour
• India censorship tracker: What the government didn’t want you to see

The post Delhi HC restrains TechWiser, TechBar from criticising AI+ smartphones without hearing them appeared first on MEDIANAMA.

The company said it paused work on chip design and foundation AI models during a business realignment in late 2025 and moved capital and staff toward AI cloud services. Krutrim now says it operates a full-stack AI cloud platform built in-house for enterprise customers.

The company reported around Rs 300 crore in revenue for FY26, nearly three times higher than the previous year, along with its first annual net profit. It said profit after tax margin crossed 10%.

“The company has reached an important milestone of being profitable, self-funded, and gaining market traction. Our AI cloud is built for Indian enterprises, by Indian engineers. The external client momentum we are seeing validates the depth of our platform,” a company spokesperson said.

From AI models to cloud services: The shift marks a sharp change from Krutrim’s earlier push into AI models and consumer AI products. Since 2024, the company has launched AI assistants, multilingual language models, translation tools, and cloud infrastructure services while positioning itself as a full-stack Indian AI company. It had also announced plans to build its own AI ecosystem, including models, cloud infrastructure, and chip-related capabilities.

However, several of Krutrim’s AI products had faced criticism over inaccurate responses, political bias concerns, and limited public technical documentation around the company’s model claims. During the beta launch of its chatbot in 2024, users had reported factual errors and misleading responses, including incorrect historical answers and confusion over the chatbot’s origin.

Weeks ago, Ola-owned Krutrim also quietly took down its AI assistant ‘Kruti’, which had been launched in 2025 as an agentic AI product for tasks such as cab booking, bill payments, food ordering, image generation, and research. The chatbot was removed from app stores and became inaccessible on the web, though the company did not publicly announce its shutdown.

Enterprise push and funding: Krutrim said it currently serves more than 25 enterprise customers across sectors, including telecom, finance, healthcare, logistics, manufacturing, and consumer internet services. It also claimed most of its GPU computing capacity has already been committed to enterprise workloads.

The pivot comes months after the Ola group moved its workloads to Krutrim Cloud and ended its use of Microsoft Azure. In May 2025, Aggarwal had posted on X: “As committed, Azure spend is now 0. All workloads on Krutrim Cloud. Within a week.”

Krutrim was launched in 2024 by Ola founder Bhavish Aggarwal and raised $50 million at a $1 billion valuation. Earlier in 2025, Aggarwal had pledged shares of Ola Electric to fund Krutrim, though the company later said those pledges were released after repayment.

The company’s latest shift also reflects the growing cost and infrastructure pressures around building advanced AI models and semiconductor systems, areas currently dominated by a small number of global firms.

Read more:

• Ola Launches its Multilingual Artificial Intelligence Model Krutrim
• Ola Krutrim Announces AI Lab Launch To Mixed Reactions From Users
• Ola Krutrim now launched for the public in its beta version: Users report inaccurate answers

The post Krutrim pivots to AI cloud, pauses chip and AI model work  appeared first on MEDIANAMA.

Sarvam AI partnered with Bengaluru-based space-tech company Pixxel to build AI data centers aboard a satellite orbiting Earth. In its blogpost, the company said that both training and inference will happen directly in orbit, without any dependence on foreign cloud or ground infrastructure.

Why does Sarvam want AI data centers in space?

• Real-time inference in orbit: “It points to a new paradigm for Earth observation, where satellites don’t just collect data for later analysis; they think for themselves and deliver conclusions,” read Pixxel’s press release. ‘Inference’ refers to the process of real-time AI responses as the user gives new input or data.
• Insights without ground lag: Establishing an AI data center in the orbital range will cut down the time period it takes for the data to reach the ground. Therefore, Sarvam says, it will enable the satellite to “flag a wildfire, track a crop disease, or monitor a pipeline leak in the same pass it observes them.” NVIDIA claims that due to orbital inference, the response time could be reduced from hours to minutes.
• To reduce foreign dependency and ground-based infrastructure: “Both training and inference happen directly in orbit, without any dependence on foreign cloud or ground infrastructure, creating a fully sovereign pipeline from observation to insight, end to end,” reads the blogpost.
  
  The company aims to achieve this through sending the 200-kg Pathfinder satellite with the GPUs that are used on the ground. The company plans to send the satellite in the fourth quarter of 2026.

What foreign companies are saying about AI data centers: Sarvam is not the first company with plans to build an AI data center in space. As of May 2026, Earth’s outer space already has at least four data centers and many more companies are planning high-powered data centers, up to gigawatt-scale. Their reasons for Orbital Data Centers (ODC) are: 

• builds resilience against terrestrial disruptions, including cyber threats, natural disasters, and geopolitical instability, says Axiom Space.
• strengthens global data sovereignty, AI autonomy, says Axiom Space
• 10x cheaper than land-based options, claims NVIDIA.
• Can use vacuum of deep space as an infinite heat sink, instead of water as a coolant, claims NVIDIA.

Cloud data centers on ground are not safe: For instance, during the ongoing Middle East conflict, in March 2026, Iran’s drones allegedly hit Amazon’s data centers in the Bahrain (UAE) region. The US tech giant said that it could take months to restore the operations.

The active AI data centerss in space:

• Kepler Communications: The US-based satellite telecommunications provider launched its first part optical relay satellites on January 11, 2026.
• Axiom Space: On the same day, along with Kepler, it also sent two orbital data center nodes.
• Zhejiang Lab: China already has 12 satellites in orbit, and after nine months of orbit-level testing, it said, “the constellation has demonstrated core capabilities, including networking, computing, model deployment and scientific payload verification,” according to China’s State Counsil of Information Office in February 2026.

Upcoming : In January 2026, SpaceX proposed to the Federal Communications Commission (FCC) an orbital data-centre constellation of up to 1 million satellites in low Earth orbit, adding to its existing fleet of over 10,000 Starlink satellites. You can read its proposal here: link.

Google’s Project Suncatcher: The tech giant aims to deploy its custom processing chips, TPUs (Tensor Processing Units), to power its machine learning in space. It aims to launch two prototype satellites by early 2027, in partnership with Planet, a US-based space tech company. Read more about it on Project Suncatcher on Google’s Research blog.

The 1960’s Outer Space Treaty doesn’t cover data provisions: “The current legal framework leaves space data centres in a grey zone: the United Nations’ Outer Space Treaty establishes no sovereignty in outer space, with launch states (a concept that presents its own issues) instead bearing responsibility and liability for space activities. Drafted in the 1960s, this treaty lacks explicit provisions regarding data,” reads a paper from the European Parliamentary Research Service. [PDF]

It further said that the European Union’s General Data Protection Regulation’s (GDPR) concept of transfers of personal data to third countries and the UN Convention against Cybercrime do not cover data transfers through satellites.

Interesting Reads: 

• The EU Space Act proposal: A GDPR for outer space? [Link]
• Orbital data centers and the legal vacuum threatening AI governance [Link]
• Who’s in the data-center space race? [Link]

Also Read: 

• OpenAI To Set Up a 1GW Data Centre in India Under the Stargate Project
• IT Ministry Seeks Cloud Services with Strict Localisation, Security Measures in AI Compute Tender
• Five Recommendations By TRAI For Regulating And Improving India’s Data Centres
• Tamil Nadu Approves Space Industrial Policy Aiming for 10,000 Jobs and Rs 10,000 Crore Investment

The post Why Sarvam wants its AI data centers in space satellites? appeared first on MEDIANAMA.

Reddit is turning to passkeys as a core defence against bots on the platform, positioning them as both a security and growth lever. Steven Huffman, Reddit’s Co-Founder and Chief Executive Officer (CEO), said during the company’s Q1FY26 earnings call that passkeys are “the lightest-weight and most privacy- and user-acceptable way of doing human verification” and “a more secure way of logging in, an easier way of logging in”. He added that the move will “help us just grow login users in general” while strengthening bot defences. “All of this work is underway,” Huffman said, adding that more features are expected this quarter.

What Reddit said about bots: Huffman framed bot control as both a transparency and safety issue. “We have what we call good bots on Reddit,” he said, referring to the tools moderators use to run communities. Reddit is shifting these to its developer platform so they can be “labeled on Reddit more transparently” while allowing the company to “batten down the hatches more on unauthorized bot usage.” He said this is “part of the human verification and defense of Reddit.”

AI and ads: Jennifer Wong, Reddit’s Chief Operating Officer (COO), said AI is central to Reddit’s ad stack. “Our strategy is to integrate more automation and AI into our ad stack,” she said, pointing to Reddit Max, which “launched to beta in early Q1”.

Wong said advertisers are seeing “a 17% reduction in cost per action and 25% more conversion outcomes” on Max campaigns, with “about 50% of Max campaign advertisers using AI-powered creative features”. 

Despite this push, Wong said Reddit is not chasing higher ad density. “Our strategy is not to increase ad load,” she said, noting it remains “quite low compared to peers”. Instead, Reddit is focused on making “the value of every impression more valuable” through machine learning, better targeting, and more advertiser competition.

AI tools: what Reddit has launched and is building

• Machine translation: Huffman said machine translation has driven recent growth, noting Reddit is “translated into 30 languages today” and can now scale further after lowering costs.
• Reddit Answers and AI search: Huffman said its AI search tool is “a great driver of retention” and “a driver of DAUq”. Reddit Answers now shows “more agentic behavior behind the scenes”, allowing comparisons of options, and is being integrated with product listings. “Search WAUqs are up 30% year-over-year,” he said.
• Post guidance and safety: Huffman described “post guidance” as “an LLM that basically helps the user navigate the rules of Reddit”. He added that Reddit is deploying “better AI-powered spam protection” while easing barriers such as age and karma limits.
• Reddit Pro for publishers: Wong said publishers can now “auto-import articles, receive AI-powered community recommendations on where to share them, and measure the reach and engagement” using Reddit Pro.
• Reddit Max expansion: Wong said Reddit is “working on putting Max in the API” and plans to onboard new advertisers directly into the system.

AI, data licensing, and Reddit’s leverage: Reddit is also positioning itself as a critical data layer for AI. Huffman said Reddit has “more than two decades of human conversation” and “over 25 billion posts and comments,” adding that “there is no artificial intelligence without actual intelligence, and that comes from Reddit.”

He said Reddit remains “the most cited source in AI citations across all platforms” and argued that demand for human input is rising as AI content proliferates. “The more it becomes sanitized and summarised and optimised for attention by AI, the more that people crave the human,” Huffman said.

On partnerships, Huffman said Reddit’s deals with Google and OpenAI are “very meaningful” and “mutual”, but indicated scope for more. “It’s citations, it’s mind share, it’s just general partnership,” he said, adding these relationships help Reddit bring in users and build its own AI tools, including machine translation and LLM-powered onboarding. However, he declined to comment on exclusivity in future deals. 

Also read

• Reddit to Label Bots, Tighten Checks on Suspicious Accounts
• Reddit Challenges Australia’s Under-16 Social Media Ban in High Court, Calls It Invalid
• Explained: Why Is Reddit Suing Perplexity AI And Other Data Scraping Companies?

The post Q1FY26 earnings call: Reddit is betting on passkeys to verify humans and grow user base appeared first on MEDIANAMA.

In a case arising from a petition filed by actress Ashu Reddy, a Telugu film actress and Bigg Boss Telugu Season 3 contestant, the Telangana High Court, in a single-judge bench of Justice P. Sam Koshy, granted interim protection against defamatory reporting. The court intervened after a trial court refused an ex parte injunction and instead issued notice. 

The petition stemmed from widespread media coverage, including videos, articles, and social media posts portraying her in connection with allegations linked to an FIR. The High Court restrained 33 respondents from publishing or disseminating defamatory material across platforms until the injunction application is decided.

What the court said:  The court placed the dispute within constitutional protections, stating that “the right to privacy is an intrinsic facet of the right to life and personal liberty guaranteed under Article 21 of the Constitution.” It then clarified the scope of this right, holding that “the right to privacy is not confined to mere secrecy but includes the right of an individual to preserve dignity, reputation, autonomy and to control dissemination of personal information.” 

Importantly, the court acknowledged the specific risks of digital publication, observing that “the injury caused by unauthorized publication of personal and private material, particularly in the digital space is immediate, pervasive and often incapable of being remedied fully by subsequent damages.”

However, the court also addressed competing claims. It noted that freedom of speech is “not absolute” and remains subject to restrictions, including defamation. Therefore, when reporting concerns about private life and it appears “prima facie defamatory and sensational”, the balance shifts. Consequently, the court found that continued publication could cause “irreversible harm” to the petitioner’s reputation and livelihood, justifying interim restraint pending adjudication.

Why this matters: The order sits within a rapidly expanding body of litigation that uses reputation to restrain digital speech. Courts are increasingly granting broad interim relief, restraining dozens of respondents simultaneously before testing the underlying merits. This trend carries real risk.

When Motorola secured a John Doe order against YouTube critics in April 2026, Apar Gupta, lawyer and founding director of the Internet Freedom Foundation (IFF), told MediaNama that the use of John Doe orders in defamation cases can create a “chilling effect” that operates even “before any creator is even named”, as the “order’s existence signals to the entire ecosystem that critical content carries legal exposure”.

Similarly, the Adani Group’s defamation suit triggered government-backed takedowns affecting publications, including Newslaundry, that were never even named as parties. The Ashu Reddy order, though narrower, follows this logic: restraint of speech first, scrutiny later. 

Also read

• Nitin Gadkari filed Rs. 50 Crore defamation suit against YouTuber over video based on Caravan report. 
• Journalist Ravi Nair Sentenced to One Year in Adani Defamation Case
• Why Supreme Court’s Observation On Decriminalising Defamation Matters For Journalism In India

The post Telangana HC grants actress Ashu Reddy protection from defamation online, cites right to privacy appeared first on MEDIANAMA.