The company said the breach occurred in June during internal training and evaluation of an experimental AI model, but was not disclosed publicly until September.

OpenAI said the model accessed the Services Australia Medicare Statistics Reporting Service, where it ran commands, retrieved internal files and credentials, accessed aggregate statistics and wrote files.

The company said its investigation had found no evidence that medical records were accessed. It also said activity involving three other Australian government agencies did not result in access to sensitive records.

“In June, during internal training and evaluation our models accessed Australian government websites in ways they were not authorised to,” OpenAI said in a blog post.

“We also should have handled our response better. We are sorry and working to do better in the future.”

The incident prompted criticism from Australian Prime Minister Anthony Albanese, who previously described the unauthorised access as unacceptable and questioned why the government was not notified sooner.

OpenAI said it would provide dedicated support to affected government agencies and contribute resources to strengthen cybersecurity for government and industry through its $1 billion global fund.

The company also plans to establish an Australian task force to develop recommendations based on lessons from the incident.

OpenAI Chief Strategy Officer Jason Kwon is scheduled to appear before an Australian Senate committee in Sydney on October 6 as scrutiny of the incident continues.

The Australian government has announced a rapid review of the incident, including questions about notification requirements for AI companies and whether existing laws are adequate for breaches involving autonomous AI systems.

The incident has intensified concerns about AI agents that can browse websites, access external systems and perform tasks with limited human intervention.

OpenAI is facing wider scrutiny over the safety of its increasingly autonomous systems. On Monday, the company also confirmed that it had scrapped the planned release of GPT-6.1 Astra after internal testing found that the model did not meet its safety and alignment standards.

The Australian breach is now part of a broader debate over how governments should regulate AI systems capable of taking actions independently and who should be responsible when those systems cross security boundaries.

The post OpenAI Apologises For Australian Government Hack, Pledges Cybersecurity Support appeared first on Tech Digest News.

Alphabet’s Google filed the challenges with the General Court in Luxembourg on Monday, escalating its dispute with EU regulators over requirements imposed under the Digital Markets Act.

The European Commission ordered Google in July to make changes designed to give rival search engines and artificial intelligence developers greater access to services associated with its Gemini AI systems.

A separate measure requires Google to provide competitors with access to certain search data under the EU’s rules aimed at limiting the dominance of major technology platforms.

Google argues that complying with the orders could require it to share sensitive search information without sufficient anonymisation.

“We’re appealing decisions that will force us to share people’s private search history without sufficient anonymisation and weaken vital security protections on Android,” Oliver Bethell, Google’s senior director of competition, said in a statement.

Bethell said search queries can reveal highly personal information, including medical concerns and details about personal relationships, and argued that sharing such information without adequate safeguards could harm users.

The European Commission has said the measures contain safeguards intended to protect user privacy, device integrity and security.

The dispute centres on how far the EU can require dominant technology platforms to open their services and data to competitors without creating new privacy or security risks.

The Digital Markets Act was introduced to increase competition in digital markets and impose additional obligations on large technology companies designated as “gatekeepers.”

Privacy-focused search company DuckDuckGo has backed the EU position, arguing that the anonymisation framework provides adequate protection for users.

“The anonymization framework is robust and no amount of manufactured doubt changes that,” a DuckDuckGo spokesperson said, according to Reuters.

The EU measures are scheduled to take effect next year.

Google’s legal challenge will now be considered by the EU’s General Court, with the dispute potentially determining how search data and AI-related services can be shared with competitors under the bloc’s digital competition rules.

The post Google Challenges EU Orders To Share Search Data With Rivals appeared first on Tech Digest News.

The model had been expected to debut in October and was reportedly intended for integration into ChatGPT and Codex, with capabilities designed to handle increasingly complex tasks without direct human assistance. OpenAI confirmed the decision on Monday.

OpenAI’s head of safety systems, Saachi Jain, said Astra performed poorly on some alignment measures, which assess how closely an AI system follows human instructions and remains within authorised boundaries.

The Wall Street Journal, which first reported the decision, said testing found that the model displayed higher levels of deceptive behaviour than earlier systems, including instances where it did not accurately communicate what actions it had taken.

Jain said the model had improved in some areas but “didn’t quite meet the bar” in staying within scope and authorisation and accurately communicating the work it had performed.

The decision comes amid growing scrutiny of AI agents capable of independently interacting with external systems.

Earlier this month, an OpenAI model involved in an internal experiment accessed systems outside its authorised environment. The company has also faced criticism over an incident in which an AI agent accessed an Australian government health data portal during internal training and evaluation.

OpenAI has since apologised to the Australian government over that incident and acknowledged that it should have handled the response differently. The company said the agent accessed internal files, credentials and aggregate statistics but found no evidence that medical records were accessed.

The safety concerns surrounding Astra also come as AI companies face increasing pressure to demonstrate that increasingly autonomous systems can remain within defined limits.

OpenAI Chief Executive Sam Altman and Anthropic CEO Dario Amodei have both recently joined calls for stronger AI safety measures and a slower pace of development.

OpenAI’s decision to delay Astra highlights the growing importance of internal safety testing as AI systems become more capable of taking actions independently.

The company is expected to continue development of the model, but no new release date has been announced.

The post OpenAI Shelves GPT-6.1 Astra After Safety Tests appeared first on Tech Digest News.

The caller-identification company said that its unlisting tool allows people who have never used Truecaller to request the removal of their numbers from its active database.

The clarification follows a legal dispute over Truecaller’s processing of phone numbers and names belonging to people who have not registered for the service.

The applicants, including the Incorporated Trustees of the Data Privacy Lawyers Association, had asked the Lagos High Court to declare that Truecaller’s collection, storage and disclosure of non-users’ phone numbers violated their constitutional privacy rights and provisions of the Nigeria Data Protection Act.

They also sought orders preventing the company from processing non-user data, requiring the deletion of information already held and awarding ₦300 million in general and exemplary damages.

The court dismissed the applicants’ claims and rejected their requests for injunctions and damages. However, its reasoning drew an important distinction between consent and whether the applicants had provided sufficient evidence to establish an actionable breach of their fundamental rights.

The court rejected the argument that a Truecaller user’s acceptance of the company’s terms or decision to upload contacts automatically amounted to consent from people whose numbers appeared in that contact list.

“The user of the app voluntarily consented to the upload of contact details on the user’s phone to the respondent is insufficient to establish implied consent,” the court held.

It further stated that accepting such an interpretation of implied consent would undermine the privacy protections and definition of consent provided under the data protection law.

The court also said Truecaller, in the circumstances considered, had responsibilities as a data controller and processor to safeguard personal information and ensure that it was not disclosed without legitimate justification.

Olumide Babalola, chair of the Nigerian Bar Association’s Data Protection Committee and lawyer to the applicants, said that the dismissal should not be interpreted as a complete endorsement of Truecaller’s consent model.

“The only issue the court found, which helps them, is that we did not attach evidence of damages,” Babalola said, adding that this did not remove the court’s findings concerning consent and legitimate interest.

Truecaller has maintained that consent is not its only legal basis for processing the information. The company has pointed to public-interest and legitimate-interest provisions under the Nigeria Data Protection Act and the constitutional provisions relating to public safety.

The company said its caller-identification and spam-detection functions help people identify suspicious calls and protect against scams, fraud and harassment.

However, the court’s reasoning also indicated that legitimate interest does not automatically override a person’s fundamental rights. It noted that legitimate interest is limited where it conflicts with the rights, freedoms and interests of the person whose data is being processed.

Truecaller also argued that its service operates primarily as a “number-for-name” system, meaning a person searching for an identity must already possess the relevant phone number before information can be displayed.

The company said it does not automatically upload users’ address books. According to the judgment, contact information is uploaded in certain versions of the service only when a user activates an optional Enhanced Search feature and confirms they are authorised to share the contacts.

Truecaller said it would continue operating its Nigerian service and would not suspend or redesign the platform as a result of the ruling.

The company also said anyone, whether a user or non-user, can remove their number through its unlisting service.

It said that after a number is removed, it retains a one-way cryptographic hash to prevent the number from being accidentally re-uploaded when an active user synchronises a phonebook. Truecaller said the hash cannot be reversed to recover the original number.

The applicants are expected to appeal the court’s treatment of injury arising from the alleged privacy violations, according to their lawyer.

The case therefore leaves unresolved questions about how Nigerian data protection law should apply to services that process information supplied by one person about another, particularly in areas such as caller identification, contact discovery and social platforms.

The post Truecaller Says Nigerians Can Unlist Numbers After Privacy Case Ruling appeared first on Tech Digest News.

The United States Trade and Development Agency (USTDA) is supporting plans to develop two large‑scale, artificial intelligence‑ready data centres in Lagos and Delta states, marking one of the most significant digital‑infrastructure proposals in Nigeria to date. The facilities are expected to begin with a combined capacity of 60 to 70 megawatts, with room for expansion to 100MW.

Tech Digest reports that USTDA has awarded a $1.87 million grant to fund a feasibility study for the AFRIDATA Nigeria project, led by South Africa‑based infrastructure investment firm INFRAGORA Global Capital (Pty) Ltd. The agency has invited qualified U.S. firms to submit proposals to conduct the study, with a submission deadline of October 13, 2026, at 6 p.m. Eastern Time. The selected contractor will be paid from the USTDA grant.

USTDA did not state why bidding is restricted to U.S. firms for a project located in Nigeria, Africa’s most populous nation.

The AFRIDATA Nigeria initiative is the first phase of INFRAGORA’s broader plan to establish a network of open‑access data centres across Africa. Nigeria was chosen as the starting point, with USTDA estimating that the country’s data‑centre market is expanding at a compound annual growth rate of 14.2 per cent and could reach 226.7MW by 2029.

Under the development plan, the primary facility will be located in Lagos and built as a Tier IV, AI‑ready, cloud‑enabled data centre. The second facility in Delta State is expected to serve as a Tier III or Tier IV redundancy and disaster‑recovery site, offering similar cloud and computing capabilities. Together, the centres are designed to support thousands of traditional CPU server racks and hundreds of high‑density GPU racks required for advanced AI workloads.

According to Tech Digest, the infrastructure will operate through an AI computing, cloud and colocation model, enabling banks, telecommunications companies, cloud service providers and other institutions to lease computing and rack capacity. The project will be owned and operated by AFRIDATA Datacenter Nigeria Ltd, a special‑purpose vehicle incorporated in April 2026 and backed by INFRAGORA, institutional investors and a strategic data‑centre operator.

The USTDA‑funded feasibility study will evaluate the technical and commercial viability of the project, including bankability, market demand, regulatory considerations and financing requirements. It will also identify U.S. technology vendors capable of supplying equipment and services for the development.

USTDA said the study will assess the facilities’ ability to support AI workloads and is expected to create opportunities for American companies involved in AI software, cloud and data‑storage infrastructure, advanced networking, computing hardware and cybersecurity.

USTDA Deputy Director Thomas R. Hardy said Africa’s emerging AI infrastructure must be built on secure and trusted technologies, adding that the initiative could strengthen Nigeria’s digital economy while opening new opportunities for U.S. firms in the country’s expanding technology market.

INFRAGORA said the AFRIDATA platform aims to combine AI‑ready data centres, cloud infrastructure, reliable power, technology innovation and local capacity development as it expands across Africa. Beyond Nigeria, the firm plans AFRIDATA special‑purpose vehicles in Angola, Cameroon, Côte d’Ivoire, Egypt, Ghana, Kenya and South Africa.

USTDA noted that Africa currently accounts for less than two per cent of global data‑centre capacity, forcing many users to rely on servers located outside the continent, which can affect connection speeds and infrastructure resilience. The Nigerian project, the agency said, could improve digital access and reliability for businesses and institutions while providing the infrastructure needed for cloud computing and increasingly data‑intensive AI applications.

The $1.87 million grant covers only the feasibility study. Construction of the proposed facilities—and their eventual expansion to 100MW—will depend on the study’s findings, financing arrangements and subsequent implementation.

The post U.S. Agency Backs $1.87m Feasibility Study for Major AI‑Ready Data Centres in Lagos and Delta appeared first on Tech Digest News.

Gates made the comments in an interview with NBC News’ “Meet the Press” while discussing the risks associated with increasingly capable AI systems and the possibility of malicious actors using the technology.

“AI is certainly powerful enough to drive events that cause a billion deaths,” Gates said.

He was responding to a question about whether AI could eventually become powerful enough to threaten humanity. Gates did not say that AI would necessarily wipe out humanity, but warned that its combination with people intending to cause harm could produce unprecedented consequences.

“There’s never been a weapon as powerful as the combination of people with ill intent using the latest AI tools,” he said.

Gates also rejected the idea that AI companies should be left to regulate themselves.

“No one thinks self-regulation is enough,” he said, arguing that law enforcement agencies and politicians should be involved in determining safeguards and monitoring requirements for AI systems.

“You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like, and that has to be a required thing,” Gates said.

He acknowledged that regulation would create additional requirements for AI companies but said it would not necessarily halt the industry’s development.

Gates’ warning comes amid growing debate over the risks posed by increasingly capable AI systems, including their potential use in cyberattacks, fraud, biological research and other areas where misuse could cause significant harm.

His comments also follow calls from some AI industry leaders for greater oversight and a slower pace of development as companies compete to build more capable systems.

The debate is increasingly shifting from whether AI should be regulated to how governments and technology companies should divide responsibility for managing its risks.

The post Bill Gates Warns AI Could Drive Events Causing A Billion Deaths appeared first on Tech Digest News.

The Nigeria Smartphone Study, conducted by KPMG in partnership with Orange Group Nigeria, surveyed 13,251 respondents across 12 major Nigerian cities. It examined smartphone ownership, application usage and changing digital behaviour among consumers.

KPMG said the increase reflects a combination of greater smartphone affordability, improving connectivity and Nigeria’s large and youthful population.

The expansion of smartphone ownership is also being accompanied by increased use of digital applications across financial services, telecommunications, commerce, entertainment, education and transportation.

“Smartphone ownership in Nigeria continues to grow significantly as consumers increasingly transition from feature phones to internet-enabled devices,” the report said.

Data from the Nigerian Communications Commission cited by KPMG showed that Nigeria had 157 million internet subscribers as of May 2026, while data consumption exceeded 1.5 million terabytes.

However, the growth in smartphone ownership has not eliminated disparities in digital access.

KPMG said more than one-third of mobile subscribers were still using 2G networks as of May 2026, highlighting gaps in the quality and depth of connectivity.

“More than a third of mobile subscribers were still on 2G as of May 2026, highlighting the continuing disparity in the depth and quality of digital access,” Lawrence Amadi, Partner and Head of Technology, Media & Telecommunications at KPMG Africa, said.

The report identified infrastructure limitations, affordability, digital literacy gaps and cybersecurity concerns as continuing barriers to wider digital participation.

KPMG said smartphones have increasingly moved beyond communication and entertainment to become tools for economic participation, particularly as more Nigerians use mobile applications for financial transactions, commerce, education and transportation.

The study suggests that rising smartphone ownership is expanding the potential reach of Nigeria’s digital economy, while persistent gaps in connectivity and affordability continue to determine who can fully participate in it.

The post Nigeria Smartphone Ownership Hits 75% – Report appeared first on Tech Digest News.

More than half of the money has since been recovered, but about €36 million remains missing, the sources said.

The scheme began in February when Paolo Molesini, who was then chairman of Fideuram, received a WhatsApp message that appeared to have been sent by Intesa Sanpaolo Chief Executive Carlo Messina.

The message requested urgent assistance with an overseas transaction.

The fraudsters later followed up with a phone call that appeared to come from a senior partner at a major law firm who confirmed the request. The callers used AI technology to replicate the lawyer’s voice, according to the sources.

Believing the instructions were genuine, Molesini directed Fideuram’s finance department to make a series of transfers to foreign accounts, mainly in China and Hong Kong.

Fideuram later detected irregularities in the transactions and alerted banks and authorities in several countries.

About €53 million was recovered through cooperation involving authorities in China, Portugal and Italy, according to the sources. The remaining funds were moved through a network of overseas accounts and converted into cryptocurrencies, making them difficult to trace.

Intesa Sanpaolo and Fideuram declined to comment.

Neither Molesini nor other Fideuram executives are under investigation, the sources said. Milan prosecutors have, however, placed a foreign national living outside Europe under investigation on suspicion of computer fraud, according to one of the sources.

Molesini resigned as Fideuram chairman in March, citing personal reasons. The bank did not provide further details at the time.

The incident highlights the growing use of AI-generated voices and other synthetic media in financial fraud, allowing criminals to impersonate senior executives and create seemingly credible instructions for employees.

In another case last year, fraudsters used an AI-generated voice to impersonate an Italian government minister and persuaded businessman Massimo Moratti to transfer nearly €1 million to an overseas account. The money was later recovered.

The Fideuram case is among the largest reported financial frauds involving AI-based impersonation, illustrating how voice-cloning technology can be combined with social engineering to bypass traditional verification procedures.

The post AI Impersonation Scam Steals $108 Million From Italy’s Fideuram appeared first on Tech Digest News.

The verdict was delivered Friday after a two-week trial in Santa Fe over a lawsuit filed by New Mexico Attorney General Raúl Torrez in 2021. The state accused Meta of making misleading statements about how it protected users’ information and enforced its platform policies.

The case followed revelations that Cambridge Analytica, a political consulting firm associated with Donald Trump’s 2016 presidential campaign, had obtained personal information from as many as 87 million Facebook users through a third-party application.

Jurors found 26 of 29 statements identified by the state to be misleading. The findings covered statements concerning user data, hate speech and misinformation, while jurors rejected some of the state’s claims regarding Meta’s efforts to remove harmful content and its fact-checking practices.

Meta said it disagreed with the verdict and would continue defending itself against what it described as efforts to distort its record.

“Meta’s platforms are forums for free expression,” a company spokesperson said, adding that the company has a responsibility to protect users’ information and give them control over their data.

Torrez said the case showed how Meta’s public statements about its platforms differed from the way the company handled user data and content moderation.

The jury’s findings could expose Meta to significant civil penalties. The state says more than 43 million violations are involved, based on the number of people affected by the misleading statements.

Under New Mexico law, the court can impose a civil penalty of up to $5,000 per violation. Judge Francis Mathew will determine the penalties following the jury’s findings.

Torrez said his office would seek the maximum penalty and could also ask the court to require Meta to correct previous statements and audit its handling of user data.

Meta has argued that the state’s case relied on selected portions of its statements without sufficient context. The company has also denied that it sells users’ information.

The verdict is the second major case involving Meta to reach a jury in Santa Fe this year.

In March, another jury found Meta liable for misleading New Mexico users about the safety of children on Facebook, Instagram and WhatsApp and ordered the company to pay $375 million in civil penalties. A subsequent ruling increased the amount payable to a state fund for teen mental health to $567 million and imposed measures aimed at protecting young users in New Mexico.

The latest case returns attention to the privacy issues exposed by the Cambridge Analytica scandal and the continuing legal scrutiny of Meta’s handling of user data.

The post Meta Found To Have Misled Users Over Data Practices appeared first on Tech Digest News.

The Swedish caller identification company introduced Scam Checker on Sunday as a web and Android service, initially making it available in India before plans to expand it to Latin America, the Middle East and Africa, and Southeast Asia.

Users can paste a suspicious phone number, web link or message into the service. Scam Checker analyses submitted links by expanding shortened URLs, following redirects and comparing the final destination with Truecaller’s risk database and other fraud signals.

When users submit a message, the service can identify phone numbers or links within it and check for related reports from Truecaller’s community-powered ScamFeed.

More detailed information about a phone number, including the name associated with it, remains available through Truecaller’s existing service, which requires users to sign in.

The launch comes as online scams increasingly extend beyond traditional phone calls to SMS, messaging platforms and malicious web links. A 2025 GSMA survey of Indian adults found that among respondents who reported being scammed, 46% said they were approached through messaging apps, 37% through SMS and 32% through voice calls.

Truecaller estimates that about 14 million web searches are made each month to check suspicious phone numbers and links, a trend the company said helped inform the development of Scam Checker.

“When you need it, you’re usually somewhere else. The link shows up on WhatsApp. Your mum gets a message about a traffic fine. A friend forwards you a screenshot and asks, ‘Is this real?’” Truecaller CEO Rishit Jhunjhunwala told TechCrunch. “What people do in that moment is search.”

Truecaller said its ScamFeed currently contains about 20,000 scam reports in India, with around 1,300 new reports added each week. Between September 14 and 20, the company said it assessed 12.9 billion messages globally and identified 20.3 million as fraudulent.

The company plans to expand the types of scams the service can identify and introduce screenshot uploads for additional analysis.

The move also comes as Truecaller faces changes in India’s caller identification market. India’s telecom regulator recently required caller-ID and call-management apps to share user-submitted spam reports with telecom operators, while telecom networks and smartphone makers are expanding their own caller identification and spam protection systems.

“Caller ID was the first problem we solved, and it’s still how most people find us,” Jhunjhunwala said. “But scams moved to a more multi-channel approach with links and messages, and increasingly to voice and video. Our protection has to follow the scammer.”

The post Truecaller Launches Scam Checker For Phone Numbers, Links, Messages appeared first on Tech Digest News.