WASHINGTON: Anthropic accidentally released part of the internal source code for its AI-powered coding assistant, Claude Code, due to &quot;human error,&quot; the company said Tuesday. An internal-use file mistakenly included in a software update pointed to an archive containing nearly 2,000 files and 500,000 lines of code, which were quickly copied to the developer platform GitHub. &quot;Earlier today, a Claude Code release included some internal source code. No sensitive customer data or credentials were involved or exposed,&quot; an Anthropic spokesperson said. &quot;This was a release packaging issue caused by human error, not a security breach.&quot; A post on X sharing a link to the leaked code had more than 29 million views early on Wednesday. The exposed code related to the tool&#39;s internal architecture but does not contain confidential data from Claude, the underlying AI model by Anthropic. Claude Code&#39;s source code was partially known, as the tool had been reverse-engineered by independent developers. An earlier version of the assistant had its source code exposed in Feb&nbsp;2025.

Artificial Intelligence

The exposed code related to the tool's internal architecture but does not contain confidential data from Claude, the underlying AI model by Anthropic.

Anthropic releases part of AI tool source code in 'error'

NEW YORK: Heavy users of artificial intelligence report being overwhelmed by trying to keep up with and on top of the technology designed to make their lives easier. Too many lines of code to analyze, armies of AI assistants to wrangle, and lengthy prompts to draft are among the laments by hard-core AI adopters. Consultants at Boston Consulting Group (BCG) have dubbed the phenomenon &quot;AI brain fry,&quot; a state of mental exhaustion stemming &quot;from the excessive use or supervision of artificial intelligence tools, pushed beyond our cognitive limits.&quot; The rise of AI agents that tend to computer tasks on demand has put users in the position of managing smart, fast digital workers rather than having to grind through jobs themselves. &quot;It&#39;s a brand-new kind of cognitive load,&quot; said Ben Wigler, co-founder of the start-up LoveMind AI. &quot;You have to really babysit these models.&quot; People experiencing AI burnout are not casually dabbling with the technology. They are creating legions of agents that need to be constantly managed, according to Tim Norton, founder of the AI integration consultancy nouvreLabs. &quot;That&#39;s what&#39;s causing the burnout,&quot; Norton wrote in an X post. However, BCG and others do not see it as a case of AI causing people to get burned out on their jobs. A BCG study of 1,488 professionals in the United States actually found a decline in burnout rates when AI took over repetitive work tasks. Coding vigilance For now, &quot;brain fry&quot; is primarily a bane for software developers given that AI agents have excelled quickly at writing computer code. &quot;The cruel irony is that AI-generated code requires more careful review than human-written code,&quot; software engineer Siddhant Khare wrote in a blog post. &quot;It is very scary to commit to hundreds of lines of AI-written code because there is a risk of security flaws or simply not understanding the entire codebase,&quot; added Adam Mackintosh, a programmer for a Canadian company. And if AI agents are not kept on course by a human, they could misunderstand an instruction and wander down an errant processing path, resulting in a business paying for wasted computing power. &nbsp;&#39;Irritable&#39; Wigler noted that the promise of hitting goals fast with AI tempts tech start-up teams already prone to long workdays to lose track of time and stay on the job even deeper into the night. &quot;There is a unique kind of reward hacking that can go on when you have productivity at the scale that encourages even later hours,&quot; Wigler said. Mackintosh recalled spending 15 consecutive hours fine-tuning around 25,000 lines of code in an application. &quot;At the end, I felt like I couldn&#39;t code anymore,&quot; he recalled. &quot;I could tell my dopamine was shot because I was irritable and didn&#39;t want to answer basic questions about my day.&quot; A musician and teacher who asked to remain anonymous spoke of struggling to put his brain &quot;on pause&quot;, instead spending evenings experimenting with AI. Nonetheless, everyone interviewed for this story expressed overall positive views of AI despite the downsides. BCG recommends in a recently published study that company leaders establish clear limits regarding employee use and supervision of AI. However, &quot;That self-care piece is not really an America workplace value,&quot; Wigler said. &quot;So, I am very skeptical as to whether or not it&#39;s going to be healthy or even high quality in the long term.&quot;

A student does an exercise during an artificial intelligence (AI) class at the College Jean Mace in Mulhouse, on March 26, 2026.

Heavy use of artificial intelligence is leading some power users to experience “AI brain fry,” a new form of mental exhaustion from managing complex digital tools.

Life with AI causing human brain ‘fry’ 

BERLIN: The artificial intelligence revolution sweeping through the entertainment sector was at first glance not evident at this year&#39;s Berlin Film Festival, but the potential for widespread changes was still on people&#39;s minds.&nbsp;The festival has had the air of an arthouse bubble when it comes to the topic of AI and the event&#39;s leadership is keeping above the fray.&nbsp;&quot;At present, we do not intend to issue any statements regarding the use of AI in the film industry,&quot; the festival said in a statement sent to AFP, adding: &quot;We are monitoring developments with great interest.&quot;&nbsp;Nevertheless, some of the filmmakers present addressed the question of how the technology is changing the industry.&nbsp;Yoshitoshi Shinomiya, director of the only animated feature in competition, &quot;A New Dawn,&quot; told reporters he had briefly considered using AI in his film.&nbsp;&quot;During production, we weren&#39;t entirely sure we would be able to complete the film. At one point we wondered whether we should use AI for the backgrounds,&quot; he said.&nbsp;But Shinomiya concluded that AI is not yet &quot;well-developed enough&quot; to do that sort of work.&nbsp;Juliette Prissard from Euro cinema, an organization representing French film and TV producers, said it&#39;s only a matter of time until the tools improve.&nbsp;&quot;It&#39;s reasonable to think that in one, two or three years... you won&#39;t be able to tell the difference anymore,&quot; she told AFP.&nbsp;AI can already &quot;write scripts&quot; and replaces extras in crowd scenes or even generate &quot;digital replicas&quot; of someone.&nbsp;No choiceIn France, where foreign-language films are frequently shown with dubbing, voice actors have already been raising the alarm about AI&#39;s impact on their profession.&nbsp;But Prissard warns other film industry jobs could be replaced in the &quot;near&quot; future, such as &quot;technicians, the set designers&quot; and even &quot;the producers themselves&quot;.&nbsp;Sevara Irgacheva, secretary general of the European Film Agency Directors&#39; association (EFAD), said that already &quot;junior jobs are disappearing: all the assistant editors, assistant screenwriters&quot;.&nbsp;Despite this, the industry &quot;is leaning toward accepting&quot; AI &quot;because, in any case, we have no choice&quot;.&nbsp;The tools have the potential to help the sector become more efficient and &quot;save time at every stage of production&quot;, particularly in the more &quot;bureaucratic&quot; aspects of the process.&nbsp;A survey carried out in early 2025 by France&#39;s National Centre for Cinema (CNC) found that 90 percent of film and audiovisual professionals surveyed were already using AI tools in their work.&nbsp;In Berlin, Austrian director Georg Tiller presented a short film mixing filmed footage and AI-generated images, saying it was an attempt to encourage his fellow filmmakers to fight for a place in the new &quot;digital cinema&quot;.&nbsp;&quot;If we don&#39;t then I fear that that we will die a slow death, because it will just steamroll over us,&quot; Tiller told AFP.&nbsp;The &#39;temptation&#39; of deregulationThe issue garnered some relief with a December agreement between OpenAI and Disney, which allowed the use of the entertainment giant&#39;s characters on Sora, the AI-generated video platform.&nbsp;In return, Disney now has &quot;privileged access&quot; to OpenAI&#39;s &quot;highly sophisticated&quot; tools, giving it a &quot;technological advantage over the rest of the sector&quot;, said Prissard.&nbsp;But the use of AI in cinema has prompted thorny legal questions over intellectual property and the very notion of authorship, at a time when legislation is only just beginning to grapple with the subject.&nbsp;Under EU rules, streaming platforms such as Netflix and Amazon Prime must carry at least 30 percent of European content in their catalogues.&nbsp;Prissard questioned how those enforcing the rules &quot;will be able to tell the difference &rdquo;Between original creations and &quot;synthetic&quot; ones.&nbsp;Given &quot;the fear of falling behind&quot; the United States and China in developing AI technologies, Prissard said that Europe may succumb to the &quot;temptation to allow more leeway to innovate without obstacles&quot;.

Film & TV

At the Berlin Film Festival, AI’s growing influence on cinema sparked debate, as filmmakers acknowledged its potential to write scripts and replace jobs, while legal and ethical concerns persist. Despite risks, most industry professionals are already adopting AI tools.

AI revolution looms over Berlin film fest

NEW DELHI: White House technology adviser Michael Kratsios said Friday that the United States &quot;totally&quot; rejects global governance of artificial intelligence. Kratsios, head of the country&#39;s delegation to an AI summit in New Delhi, made the comments ahead of an expected leaders&#39; statement setting out a shared vision on how to handle the divisive technology. &quot;As the Trump Administration has now said many times: We totally reject global governance of AI,&quot; he said at the summit, which draws to a close on Friday. &quot;AI adoption cannot lead to a brighter future if it is subject to bureaucracies and centralised control.&quot; On Friday morning, UN chief Antonio Guterres had said that a new expert panel convened by the global body aimed to &quot;make human control a technical reality&quot;. The advisory group -- aiming to be to AI what the UN&#39;s Intergovernmental Panel on Climate Change (IPCC) is to global warming -- was created in August, and its 40 members have now been confirmed, Guterres said. At last year&#39;s edition in Paris, US Vice President JD Vance warned against &quot;excessive regulation&quot; that &quot;could kill a transformative sector&quot;. In New Delhi, Kratsios said that &quot;international discussion of AI has evolved, as this summit itself attests,&quot; noting the change of the meeting&#39;s name from &quot;AI Safety&quot; to &quot;AI Impact&quot;. &quot;This is clearly a positive development,&quot; but &quot;too many international forums, such as the UN&#39;s Global Dialogue on AI Governance, maintain a general atmosphere of fear,&quot; he said. &quot;We must replace that fear with hope,&quot; Kratsios added, saying that AI has the potential to &quot;advance human flourishing and drive unprecedented prosperity&quot;. He argued that &quot;ideological, risk-focused obsessions, such as climate or equity, become excuses for bureaucratic management and centralisation&quot;. &quot;In the name of safety, they increase the danger that these tools will be used for tyrannical control.&quot; &quot;Focusing AI policy on safety and speculative risks... inhibits a competitive ecosystem, entrenches incumbents, and isolates developing countries from full participation in the AI economy,&quot; Kratsios said.

White House technology adviser Michael Kratsios (File/World Economic Forum)

Americas

White House tech adviser Michael Kratsios opposed centralized oversight, urging innovation over regulation, while UN chief Antonio Guterres promoted an expert panel to ensure human control of the technology.

US 'totally' rejects global AI governance: White House adviser

NEW DELHI: Sam Altman, head of ChatGPT maker OpenAI, told a global artificial intelligence conference on Thursday that the world &quot;urgently&quot; needs to regulate the fast-evolving technology. An organization could be set up to coordinate these efforts, similar to the International Atomic Energy Agency (IAEA), he said. Altman is one of a host of top tech CEOs in New Delhi for the AI Impact Summit, the fourth annual global meeting on how to handle advanced computing power. &quot;For an obvious example, there&#39;ll be extremely capable biomodels available open-source that could help people create new pathogens, he said on stage. &quot;We need a society-wide approach about how we&#39;re going to defend against this.&quot; Frenzied demand for generative AI has turbocharged profits for many companies while fuelling anxiety about the risks to individuals and the planet. &quot;Democratisation of AI is the best way to ensure humanity flourishes,&quot; Altman said, adding that &quot;centralisation of this technology in one company or country could lead to ruin&quot;. &quot;This is not to suggest that we won&#39;t need any regulation or safeguards,&quot; he said. &quot;We obviously do, urgently, like we have for other powerful technologies.&quot; Many researchers and campaigners say stronger action is needed to combat emerging issues, ranging from job disruption to sexualised deepfakes and AI-enabled online scams. &quot;We expect the world may need something like the IAEA for international coordination of AI,&quot; with the ability to &quot;rapidly respond to changing circumstances&quot;, Altman said. &quot;The next few years will test global society as this technology continues to improve at a rapid pace. We can choose to either empower people or concentrate power,&quot; he added. &quot;Technology always disrupts jobs; we always find new and better things to do.&quot;

OpenAI CEO Sam Altman attends the AI Impact Summit in New Delhi on February 19, 2026. (Photo by Ludovic MARIN / AFP)

Sam Altman urges global coordination and safeguards for advanced AI, warning of biosecurity risks, power concentration, and rapid societal disruption.

OpenAI's Altman says world 'urgently' needs AI regulation

Web Desk

ISLAMABAD: Discrimination against Muslim and Christian minorities in India is being fuelled by two converging trends, the viral spread of generative AI hate content and the rapid expansion of AI-enabled policing and surveillance, warns a report published on Feb. 11 by the Centre for the Study of Organized Hate&nbsp;(CSOH). The report has come ahead of India AI impact summit, 2026, which will be from Feb. 16 to 20. The CSOH report says hate narratives that have circulated for years; often packaged as conspiracies such as &ldquo;love jihad&rdquo; and &ldquo;population jihad,&rdquo; or fuelled through cow-protection and temple-mosque disputes, are now being amplified with synthetic media that looks real.&nbsp;Text-to-image and text-to-video tools can produce photorealistic depictions, caricatures, and &ldquo;evidence-like&rdquo; scenes that reinforce stereotypes, Muslim men framed as violent or criminal, and Muslim women dehumanized through sexualized, non-consensual imagery. Communal targetingMoments of public tragedy are turned into accelerants for communal targeting, with AI-generated visuals circulated alongside inflammatory commentary to cast minorities as villains in the immediate aftermath of attacks or accidents. The report highlights that political messaging has also begun using generative AI directly.&nbsp;Official social media accounts linked to the&nbsp;ruling&nbsp;BJP, have posts that depict Muslims as &ldquo;ghuspaithye/infiltrators&rdquo; and portray opposition leaders as &quot;secretly&quot; aligned with them.&nbsp;A pattern of dehumanizing metaphors is developing, where minorities are rendered as pests or threats, paired with meme-like humor or dramatic background music that helps normalize extreme hostility, especially in battleground states in the run-up to 2026 assembly election such as Assam and Delhi. A second cluster of concerns centers on the state&rsquo;s deployment of AI. Predictive policingIt notes remarks by Devendra Fadnavis, CM Maharashtra,&nbsp;about an AI tool being developed with IIT Bombay to detect alleged undocumented migrants using language cues, an approach that risks profiling Bengali-speaking Muslims and low-income migrant workers.&nbsp;The spread of predictive policing has also been flagged, large criminal databases and facial recognition are systems that can embed historical bias into &ldquo;data-driven&rdquo; enforcement. Examples of this &ldquo;data-driven&rdquo; enforcement referenced in the report include long-running predictive systems used by Delhi Police and expanding surveillance networks in cities such as Hyderabad, Bengaluru, and Lucknow.&nbsp;Without transparent safeguards, independent oversight, and clear legal limits, AI risks hardening discrimination into automated, scalable practice, the report states.

India

Generative AI and predictive policing are fuelling discrimination against India's minorities through deepfake hate speech and automated surveillance. A report from the Centre for the Study of Organized Hate (CSOH) warns of these issues.

AI enabled hate on the rise in India: Report

Sam Altman said OpenAI hired OpenClaw creator Peter Steinberger, plans multi-agent “personal agents” as core products, will keep OpenClaw open-source, amid hype, security concerns, and pressure to monetize ChatGPT via ads.

OpenAI hires creator of 'OpenClaw' AI agent tool

ISLAMABAD: Prime Minister Shehbaz Sharif announced that Pakistan would invest $1 billion in artificial intelligence by 2030. He was speaking at the inaugural ceremony of Indus AI Week in Islamabad on Monday. Addressing participants at the event organized by the Ministry of Information Technology and Telecommunications, the prime minister said the investment would go a long way toward building indigenous AI infrastructure in the country. &quot;[This] event will not only transform Pakistan&rsquo;s technological landscape but will also prove to be a real game changer. In collaboration with our friendly and brotherly countries, we will move forward on this shared path with strong commitment and enthusiasm,&rdquo; he said. The PM said that AI would be integrated into the federal school curriculum across the country, including the tribal districts of Khyber Pakhtunkhwa, Azad Jammu and Kashmir, Gilgit-Baltistan, and the remote parts of Balochistan. &ldquo;This is to prepare the youth for leadership and the digital economy,&rdquo; he said.&nbsp; He announced that the government will provide 1,000 fully funded PhD scholarships in AI to youth by 2030, with the aim of building a world-class research center. PM Sharif said the government will launch a nationwide program to train one million non-IT professionals in AI skills, enabling them to improve productivity and their livelihoods. He said Pakistan was fully ready to join the global AI community, adding that it would be a valuable partner in this field. The prime minister said he will implement programs to transform the country&rsquo;s youth from IT technicians into AI experts. Highlighting priority sectors, the prime minister said AI interventions would focus on agriculture, mines and minerals, industry, commerce, trade, and youth empowerment.&nbsp; With a population of around 240 million, nearly 60% of whom are young, he stressed the need to equip them with modern knowledge and skills.&nbsp; Minister for IT Shaza Fatima Khawaja said Indus AI Week aimed to strengthen coordination between universities, governments and international companies, underscoring Pakistan&rsquo;s full commitment to a digital revolution. She said the Pakistan Digital Authority was preparing a nationwide digital master plan to guide future transformation. Minister for Planning Ahsan Iqbal, in his remarks, said the world had entered a moment where intelligence itself had become a factor of production, with nations competing on ideas, talent, data and technology rather than commodities.&nbsp; He described artificial intelligence as a bigger disruptor than electricity or the internet, noting that Pakistan&rsquo;s engagement with technology began over two decades ago with early IT policies, the creation of NADRA, and large-scale investment in advanced human capital.

Economy

He said the investment would go a long way toward building indigenous AI infrastructure in the country.

Indus AI Week: Pakistan PM announces $1B AI investment by 2030

PARIS: Google unveiled an artificial intelligence tool Wednesday that its scientists said would help unravel the mysteries of the human genome and could one day lead to new treatments for diseases.&nbsp;The deep learning model AlphaGenome was hailed by outside researchers as a &quot;breakthrough&quot; that would let scientists&rsquo; study and even simulate the roots of difficult-to-treat genetic diseases.&nbsp;While the first complete map of the human genome in 2003 &quot;gave us the book of life, reading it remained a challenge&quot;, Pushmeet Kohli, vice president of research at Google DeepMind, told journalists.&nbsp;&quot;We have the text,&quot; he said, which is a sequence of three billion nucleotide pairs represented by the letters A, T, C and G that make up DNA.&nbsp;However, &quot;understanding the grammar of this genome -- what is encoded in our DNA and how it governs life -- is the next critical frontier for research,&quot; said Kohli, co-author of a new study in the journal Nature.&nbsp;Only around two percent of our DNA contains instructions for making proteins, which are the molecules that build and run the body.&nbsp;The other 98 percent was long dismissed as &quot;junk DNA&quot; as scientists struggled to understand what it was for.&nbsp;However, this &quot;non-coding DNA&quot; is now believed to act like a conductor, directing how genetic information works in each of our cells.&nbsp;These sequences also contain many variants that have been associated with diseases. It is these sequences that AlphaGenome is aiming to understand.&nbsp;A million lettersThe project is just one part of Google&#39;s AI-powered scientific work, which also includes AlphaFold, the winner of 2024&#39;s chemistry Nobel.&nbsp;AlphaGenome&#39;s model was trained on data from public projects that measured non-coding DNA across hundreds of different cell and tissue types in humans and mice.&nbsp;The tool is able to analyze long DNA sequences then predict how each nucleotide pair will influence different biological processes within the cell.&nbsp;This includes whether genes start and stop and how much RNA -- molecules which transmit genetic instructions inside cells -- is produced.&nbsp;Other models already exist that have a similar aim. However, they have to compromise, either by analyzing far shorter DNA sequences or decreasing how detailed their predictions are, known as resolution.&nbsp;DeepMind scientist and lead study author Ziga Avsec said that long sequences -- up to a million DNA letters long -- were &quot;required to understand the full regulatory environment of a single gene&quot;.&nbsp;And the high resolution of the model allows scientists to study the impact of genetic variants by comparing the differences between mutated and non-mutated sequences.&nbsp;&quot;AlphaGenome can accelerate our understanding of the genome by helping to map where the functional elements are and what their roles are on a molecular level,&quot; study co-author Natasha Latysheva said.&nbsp;The model has already been tested by 3,000 scientists across 160 countries and is open for anyone to use for non-commercial reasons, Google said.&nbsp;&quot;We hope researchers will extend it with more data,&quot; Kohli added.&nbsp;&#39;Breakthrough&#39;&nbsp;Ben Lehner, a researcher at Cambridge University who was not involved in developing AlphaGenome but did test it, said the model &quot;does indeed perform very well&quot;.&nbsp;&quot;Identifying the precise differences in our genomes that make us more or less likely to develop thousands of diseases is a key step towards developing better therapeutics,&quot; he explained.&nbsp;However, AlphaGenome &quot;is far from perfect and there is still a lot of work to do&quot;, he added.&nbsp;&quot;AI models are only as good as the data used to train them&quot; and the existing data is not very suitable, he said.&nbsp;Robert Goldstone, head of genomics at the UK&#39;s Francis Crick Institute, cautioned that AlphaGenome was &quot;not a magic bullet for all biological questions&quot;.&nbsp;This was partly because &quot;gene expression is influenced by complex environmental factors that the model cannot see&quot;, he said.&nbsp;However, the tool still represented a &quot;breakthrough&quot; that would allow scientists to &quot;study and simulate the genetic roots of complex disease&quot;, Goldstone added.

Google’s AI, AlphaGenome, studies DNA, especially “non-coding” parts, to see how genes work and how changes can cause disease. It helps scientists understand genetics better, though it can’t account for all environmental effects.

Google unveils AI tool probing mysteries of human genome

BRUSSELS: Google must provide rival AI services equal access to its features and other search engine platforms access to data, the EU said on Tuesday, as it said it would help the giant over six months to comply with rules.&nbsp;The EU executive said it would launch proceedings to help <a href="https://www.pakistantv.com/article/eu-launches-antitrust-probe-into-googles-data-use-for-ai" target="_blank" style="color:#0073e6;" rel="noopener noreferrer">Google</a> prepare measures in line with its flagship rulebook, the Digital Markets Act (DMA).&nbsp;Under the DMA, the world&#39;s biggest tech companies must open up to competition to give consumers more options and limit abuses linked to market dominance.&nbsp;US President Donald Trump&#39;s government has railed against the law and its sister content moderation law the Digital Services Act, accusing Brussels of unfairly targeting US firms.&nbsp;The European Commission wants to ensure Google gives rival AI service providers equal access to its Android operating system, and demands the American titan grants competing search engines access to search data.&nbsp;Brussels believes the move will allow rivals &quot;to optimise their services and offer users genuine alternatives to Google Search&quot;.&nbsp;The EU step is not a formal investigation that could lead to fines.&nbsp;But if Brussels is not satisfied with Google&#39;s efforts, it can later conclude the company is not complying.&nbsp;And any DMA violations can lead to fines of up to 10 percent of a company&#39;s total global turnover.&nbsp;&quot;We want to help Google by explaining in more detail how it should comply with its interoperability and online search data sharing obligations under the Digital Markets Act,&quot; EU competition chief Teresa Ribera said in a statement.&nbsp;Google pushed back, insisting Android is open by design.&nbsp;&quot;We&#39;re already licensing search data to competitors under the DMA,&quot; Google&#39;s senior competition counsel Clare Kelly said in a statement.&nbsp;&quot;However, we are concerned that further rules which are often driven by competitor grievances rather than the interest of consumers, will compromise user privacy, security, and innovation.&quot;&nbsp;The commission said it would conclude the proceedings within six months.&nbsp;Google already faces a fine under the DMA for favouring its own services in a probe launched in March 2024.&nbsp;The giant has also been subject to a separate DMA probe since November over suspicions it pushed down news outlets in search results.&nbsp;The scrutiny does not end with the DMA. Google also faces cases under the bloc&#39;s competition rules.&nbsp;In December, the EU said it had opened a probe to assess whether Google breached antitrust rules by using content put online by media and other publishers to train and provide AI services without appropriate compensation.&nbsp;That came after the EU slapped Google with a 2.95-billion-euro ($3.5 billion) fine in September for breaking competition rules.

The EU told Google to give rival AI and search engines equal access under the Digital Markets Act, launching a six-month compliance process. Non-compliance could trigger major fines, as Google already faces multiple EU probes.

Articles tagged with "AI"