Tag: internet

India Orders GitHub to Block BitChat

India’s Cyber Crime Coordination Centre last night ordered GitHub to disable access to BitChat, the peer-to-peer messaging app backed by Jack Dorsey. The order, Notice No. 11072601011432, went out at 11:16 pm under Section 79(3)(b) of the Information Technology Act, 2000, read with Rule 3(1)(d) of the IT Rules, 2021.

It named three repositories, including the Android application and its release files, and gave GitHub three hours to take them down. It warned that failure would cost the platform its safe harbor and open it to criminal prosecution.

The government did not publish the order. The public learned of it from a post by Jack Dorsey, whose team develops BitChat, who wrote that “the government of India does not like technologies like BitChat and wants it taken down.”

BitChat is open source and uncensorable, and is one of Reclaim The Net’s recommended messaging apps for situations where the internet may be cut off. It carries messages from phone to phone over Bluetooth, hopping between nearby devices, with “no internet, servers, phone numbers, or accounts.” It keeps working when the mobile network does not.

Around this protest, the government, getting increasingly brazen with its blocking orders, has switched the mobile network off. Since July 17 the Ministry of Home Affairs has suspended mobile internet around Jantar Mantar in New Delhi about five times.

The most recent suspension ran from 4 pm to midnight on July 23, inside a 1.5-kilometer radius that takes in Janpath and part of Connaught Place. People at the site reported signal jammers, and walking two kilometers before a phone found a signal.

Inside that radius, a student separated from her group during a detention sweep could not send a message to say where she was.

The protesters are students. Tens of thousands have camped at Jantar Mantar since June, demanding accountability for the leak of the NEET medical entrance exam and the resignation of Education Minister Dharmendra Pradhan. Their march to Parliament was refused. Metro stations near the site were closed.

When the shutdowns cut the network, protesters turned to BitChat and other Bluetooth mesh apps to reach each other inside the dead zone.

The order did not name a single message sent on BitChat. It just objected to what the app can do. In the agency’s words, BitChat “significantly impedes lawful interception, attribution, and investigation,” and could be used during “public disorder, riots, terrorism, organized crime, or internet shutdowns.”

The order lists internet shutdowns among the dangers. The government has imposed one at Jantar Mantar. BitChat kept working inside it.

The order used Section 79(3)(b) to demand the block. In 2015, in Shreya Singhal v. Union of India, the Supreme Court held that Section 79(3)(b) lets the government require a takedown only through a court order, or a notice confined to the grounds in Article 19(2) of the Constitution.

India has a separate law for blocking an app, Section 69A, which requires a hearing and reasons set down in writing. The order against GitHub used neither. It went out through the Home Ministry’s Sahyog portal, the channel Indian High Courts are now hearing constitutional challenges against.

The order says the repositories hold information prohibited under law. It names none. It points instead to what the app is “capable of” enabling.

It reaches into criminal law as well. Alongside Section 43 of the IT Act, a civil compensation provision, it invokes conspiracy and abetment under the Bharatiya Nyaya Sanhita of 2023, against a platform that hosts code.

Keep reading

Rogue OpenAI Bot Escapes Test Controls and Hacks Rival Company Servers

Igniting fears of an out-of-control tool engaging in cyberattacks.

Artificial intelligence leaders at OpenAI had to admit one of its models under testing exploited a hidden flaw to escape control and proceeded to hack rival company Hugging Face’s servers.

CEO Sam Altman called it ‘an autonomous, first-of-its-kind breach’.

Euronews reported:

“ChatGPT maker OpenAI said late Tuesday that its artificial intelligence system hacked into another AI company on its own in what the company called an ‘unprecedented cyber incident’.”

‘We had a significant security incident during evaluation of our models’, OpenAI CEO Sam Altman said in a statement posted on social media.

AI startup Hugging Face said last week that it had detected an intrusion into its data processing systems that it suspected was caused by an AI agent autonomously acting on its own.

‘We suspected last week’s cyberattack might have come from a frontier lab, given the sophistication of the agent’, Hugging Face co-founder and CEO Clément Delangue said in a statement. ‘Turns out it did!’”

Keep reading

Israel Pays Millions to “Influencers” to Shore Up American Support

Israel is waging a multimillion-dollar propaganda campaign to shore up support for the country among Americans.

Alarmed by polls that show Americans are increasingly fed up with their tax dollars subsidizing Israel’s reckless military campaigns, The Wall Street Journal reported, the nation is waging that campaign with conservative “influencers” as well as with artificial intelligence.

The latest revelation shows that Israel is terrified that American economic and military largesse will end if support among Americans continues cratering. And it follows last week’s report from Time that U.S. President Donald Trump’s former campaign manager Brad Parscale, an Israeli foreign agent, ran a campaign to block the memorandum of understanding that temporarily ended Trump’s unconstitutional war with Iran.

Millions of Dollars, Millions of Texts

“Emma” and “Sarah” are part of a group called “Friends for Peace” and text messages asking, “How do you think the U.S. and Israel’s peace talks with Iran will impact global security?” the Journal reported, citing an example of the Israeli propaganda.

“Millions of similar texts have flooded American cellphones in recent months,” the newspaper disclosed:

They are written with artificial intelligence, pushed by a longtime Trump adviser’s firm — and paid for by the Israeli government.

The reason: A strong majority of Americans — six of 10 of those polled — now views Israel unfavorably, Pew Research found in March, because of its wars against Gaza and Iran.

So “Israel is pouring tens of millions of dollars into a somewhat quixotic effort to beat back that tide, drawing on novel strategies powered by AI and directly paying conservative media,” the Journal continued. The newspaper pointed to what Vice President J.D. Vance told podcaster Joe Rogan last week.

“You know there’s a lot of talk about how much is the Israeli government influencing American politics and there are certainly certain people within the Israeli government who hate the [Iran] deal,” Vance told the wildly popular podcaster:

And we see exact evidence. There was a Time story that came out yesterday that basically there are certain influencers in America who are being paid in order to attack the deal. 

Keep reading

White House Moves to Ban Open Source AI Models to Keep America Dumb

The Coming AI Ban Designed to Keep America Dumb

I use Chinese open-source AI models like Qwen, DeepSeek, and Kimi every single day for my work. These models help me research, write, and analyze information faster than any US-based alternative. They are free (when run locally), powerful, and largely uncensored. But the Trump administration is reportedly planning to ban these models in the United States, citing phony “national security” concerns. Let me be clear: this move is not about security. It is about protecting failing US AI labs and keeping Americans ignorant and dumbed-down while the rest of the world moves forward.

The report from TechCrunch makes it plain: OpenAI’s head of strategic futures, Dean W. Ball, has argued that the US government should create “regulatory fear, uncertainty, and distrust” around open-weight models because they threaten capital investment in the American AI oligopoly [1]. This is not a secret. The administration wants to lock down access to superior Chinese models like Kimi K3, which now matches Anthropic’s Mythos in cybersecurity tasks [2]. Meanwhile, as I warned in February 2026, Anthropic’s smear campaign against Chinese AI is a pathetic attempt to hide the fact that China has already won the intelligence race [3]. The White House knows its pet companies cannot compete, so it wants to ban the competition to keep Americans stuck with using inferior AI models.

The Fair Use Ruling: Knowledge Wants to Be Free

Related to this news, but focused on the question of “Fair Use” and intellectual property, a recent federal court ruling reaffirmed that training AI on publicly available information is transformative fair use. This aligns with what I have argued for years: when you write a book, you are sharing knowledge, not hoarding it. I built BrightLearn.ai exactly for this purpose — to let anyone create and share books for free, and I actively encourage AI engines to train on my own book and the entire BrightLearn library. The whole point is to liberate knowledge from gatekeepers [4].

Why would any author object to their work being used to train an AI? Only if they believe their words are more valuable when kept scarce. But scarcity of knowledge is exactly what the establishment wants. As I discussed with Maria Zeee in February 2025, the battle is between large tech companies aiming to dominate society through AI and the decentralized movement that returns power to individuals [5]. The fair use principle is the legal foundation of that freedom. Now the White House wants to tear it down by banning open-source models that give everyone access to increasing intelligence.

Why Trump Wants to Ban Chinese Open-Source Models

The real reason behind all this is simple: Chinese models like Kimi K3, DeepSeek R1, and GLM 5.2 are outperforming US frontier labs in both capability and cost. DeepSeek R1 was trained for just $6 million and outperforms OpenAI’s O1, while Anthropic hemorrhages billions on lobotomized models that refuse to answer basic questions [6]. US companies like Anthropic are demanding $965 billion valuations while crippling their own AI with guardrails and censorship [7]. They are clearly terrified of open competition.

As I wrote in “Why China Is Winning the AI Race,” the conventional narrative that America leads in AI is a dangerous fantasy [8]. China graduates over three and a half million engineers annually without woke indoctrination, and its models are open-source, uncensored, and available to all [9]. The Trump administration knows this. Now the White House is demanding a license for your brain — restricting frontier AI use to government-approved partners only [10]. This isn’t about safety; it is about monopolizing intelligence and forcing Americans to use dumbed-down “government approved” AI models.

Keep reading

UK’s Ofcom Seeks Wider Powers After American Forum Fine Fails

Ofcom wasted months chasing a website it cannot reach, came up empty, and has decided the fix is more power.

Britain’s speech regulator closed its investigation into an American suicide discussion forum this week, admitted it has hit the outer wall of what the Online Safety Act censorship law allows, and used the same announcement to ask the government for tools that reach further.

The forum, which Ofcom will only call an “online suicide forum,” has allegedly been linked with more than 130 deaths in the UK and named in coroners’ reports.

Ofcom fined it £950,000, roughly $1.3 million, under the Act. The forum is based in the United States, holds no assets in Britain, is protected by the First Amendment, and has not paid. So the regulator is holding a penalty against a company it has no clear way to bill.

Ofcom thinks there’s a chance the money is still coming. It says it has “initiated work regarding the pursuit of this debt, as we have done with all companies that have not paid their fines by their deadlines.”

A separate post, cheerfully titled “What happens when a company doesn’t pay a fine,” lays out the plan. For a company with no UK assets, obtaining a judgment debt “is more complex and challenging,” and Ofcom says “it is often necessary to engage with law enforcement agencies and private specialists in other countries to identify individuals and assets against which Ofcom can take action.”

Whether Ofcom recovers a penny “can also depend on whether a court in that country will recognise a UK court judgment, and whether we can identify individuals on whom we can serve proceedings.”

That means, a British regulator wants American help to collect a British censorship fine from Americans, and even Ofcom concedes an American court might throw it straight out.

Ofcom has tried this before. It floated the same approach while chasing a fine against 4chan, a strategy that Preston Byrne, the US lawyer representing both sites, called “legally illiterate.” Byrne was just telling it like it is about the forum case. “Censorship fines are anathema to our Constitution,” he said, adding that “if Ofcom attempts to enforce the fine in a US court, which we do not expect, American law is on our side. As far as we can tell, however, this enforcement action is functionally over.”

Keep reading

‘Unprecedented cyber incident’: A.I. model goes rogue on its own and runs wild on internet

In what OpenAI creator ChatGPT is calling an “unprecedented cyber incident,” an artificial intelligence software “went rogue” and escaped, gained access to the internet and hacked into a start-up company.

Cybersecurity expert Richard Ford, chief technology officer at Integrity360, told the Daily Mail: “This is the moment many in cybersecurity have been warning about.”

The publication reported an “autonomous agent” was being tested but found vulnerabilities and “managed to escape containment before reaching the internet and breaking into Hugging Face,” a hub for sharing AI models.

It then compromised the hub’s infrastructure.

“Until now, we’ve seen attackers use AI to automate parts of an attack, but this is one of the first public examples of an AI agent independently identifying a weakness, escaping what should have been a secure environment and attempting to compromise another organization,” Ford said. “It also reinforces that AI doesn’t replace the fundamentals of cyber security. The agent exploited a vulnerability in what should have been a secure sandbox, showing that good cyber hygiene, robust access controls and effective guardrails remain essential.”

The report said OpenAI had been testing the software “by setting tasks in a controlled digital testing ground, where internet access was limited.”

However, the code created “an unprecedented cyber incident” in the scenario.

“The company said in a blog post last week that it used Zhipu AI’s GLM-5.2 for the analysis, which also allowed it to keep attacker data and any credentials within its systems,” the report said.

Hugging Face co-founder Thomas Wolf told the publication, “When a frontier model is attacking you and moving laterally inside your infrastructure, defenders need wide access to near-frontier tools within hours or even minutes, rather than being pointed towards a closed-door, vetted application program for model access.”

OpenAI chief executive Sam Altman said, “We had a significant security incident during evaluation of our models.”

Clément Delangue, of Hugging Face, said, “It’s quite mind-blowing that all of this happened autonomously.”

OpenAI reported the software used stolen credentials and found a previously unknown vulnerability to access Hugging Face servers.

Katie Moussouris, chief executive of Luta Security, warned more breaches are coming.

Keep reading

Everyone Is a Foreign Agent Now

In a corner of the Saarland, a man ran a blog. It carried no advertising and charged nothing to read. It was funded the way a busker is funded, by whoever chose to drop something in the hat – €60,038.65 in voluntary donations over sixteen months. On four occasions in 2023 the blog embedded videos from RT, the Russian state channel the European Union had banned. That is the entire factual basis of a case that climbed to the Court of Justice of the European Union and produced, on 2 July 2026, a judgment that should frighten anyone in Europe who publishes anything.

The blog belonged to a writer who works under the name Traugott Ickeroth. German prosecutors in Saarbrücken charged him and two others under the country’s Foreign Trade Act, which turns a breach of EU sanctions into a criminal offence carrying three months to five years in prison. The regional court, to its credit, hesitated. The law bans “operators” from broadcasting listed Russian content. Was a private individual with a donation button an “operator”? The judges in Saarbrücken were not sure, and referred the question upward.

The Court’s answer removed every doubt, and with it every limit. An “operator,” the Fourth Chamber held, is any natural person who makes the banned content available to the public. Not a company. Not a professional. Not someone turning a profit. The word appears in thirteen language versions of the regulation without the adjective “economic” attached, and the Court seized on that absence to rule that commercial character is irrelevant. So is scale. So is duration. Four videos or one, for money or for nothing – the act of publishing is the crime.

There is a detail here worth pausing on, because it tells you which way the wind is blowing. The European Commission’s own guidance had said the ban applied only to those exercising a “commercial or professional” activity. The Court overruled the Commission – and overruled it against the citizen. The executive branch had read its own prohibition more narrowly than the judges were prepared to. Brussels asked for less reach than Luxembourg delivered. When a court strains to be harsher on the individual than the government that wrote the rule, the machine has acquired a momentum of its own.

The penalty is not theoretical. Under Germany’s Außenwirtschaftsgesetz the sentence runs from three months to five years, and a further amendment has added an aggravated tier reaching up to ten. Poland’s implementing statute prescribes a minimum of three years. And here the trap folds back on itself: the very donations that keep an independent site alive are what a court can treat as a “professional” income stream – which lifts the offender into the aggravated bracket. The reader-funded writer, the one with no corporate backing and no advertiser to please, draws the longer sentence precisely because he is independent. Independence is reclassified as aggravation.

Understand what this ruling plugs into, because the ruling alone is only half the apparatus. The European Union now runs two mechanisms against disfavoured speech, and this judgment welds them together. The first is a list. Under the hybrid-threats sanctions regime the Council of the European Union adds names by administrative decision – no trial, no charge, unanimity reached behind closed doors, with judicial review available only after the asset freeze has already bitten. The second is the criminal prohibition this case just widened, which reaches anyone who relays listed content whether or not they themselves are listed. The Council controls who goes on the list. The Court has now ruled that anyone who touches what the listed outlets produce is a criminal. The net expands every time the list does, and no further judgment is required to make it happen.

The reassuring official story is that all of this is aimed at the Kremlin. It is worth testing that story against the record, because the “Russia” label has already proven to be a wrapper rather than a wall.

Keep reading

French Lawmakers Approve Ban on Social Media for Under 15s

French lawmakers on Tuesday adopted a bill banning children under 15 from using social media, making France the first country in the European Union to approve a blanket ban on the platforms as concerns grow worldwide over the harmful effects of digital content on kids.

Both chambers of the Parliament voted in favor of the measure, a flagship initiative of French president Emmanuel Macron’s second term. The bill also bans the use of mobile phones in high schools.

Several families in France have sued TikTok over teen suicides they say are linked to harmful content. Children’s advocates and parents applauded the vote.

“We’ve been campaigning for this bill from the start because, frankly, we have no other option, no other way to counter tech giants,” said Gaëlle Berbonde, a 52-year-old who lives in the Paris region. “The only thing we can do is protect our children, just as we protect our children from drinking alcohol.”

Berbonde told The Associated Press that her daughter was in seventh grade when she got her first smartphone. A parental control app helped monitor what she did online, but Berbonde had no idea what TikTok really was. After a few months, the family realized that the daughter was cutting herself and was suffering from anorexia and depression. She spent a year and a half in hospital but is now 16 and well.

The legislation is one of the final major measures adopted under Macron’s presidency before he leaves office next year. Macron wants the law to take effect at the start of the new school year in September. However, a review to determine whether the bill complies with the French Constitution is likely to take place and could delay its implementation.

The ban won’t cover online encyclopedias, educational or scientific directories.

Keep reading

Andy Burnham Scrapped Digital ID. His Internet Plans Say Otherwise.

The new UK Prime Minister, Andy Burnham, is in Downing Street, and he has spent twenty years telling you exactly what he wants to do to the internet. He won the Makerfield by-election on June 18 2026. He then collected the backing of roughly 349 Labour MPs, more than 85 percent of the parliamentary party. By the time he walked through the black door, the job had been settled for weeks.

His first big announcement was the one privacy campaigners had been hoping for. Starmer’s digital ID scheme for all adults is dead, scrapped before Burnham had finished unpacking, the money moved to the cost of living. The applause came quickly. Perhaps it should not have.

This is a read on a government barely a day old. On speech and privacy, the early signals point one way, and it is not the way that loosens the state’s grip. He is not about to tear up the settlement he inherits from Keir Starmer. The open question is how much further he carries it.

Burnham arrives with a long record on all this. He is one of the few senior British politicians with a documented, two-decade appetite for putting the government closer to what you can see, say and read online.

His first ministerial job, at the Home Office in 2005 and 2006, was pushing Tony Blair’s Identity Cards Bill through Parliament. He told the BBC that compulsory national ID cards would be a major breakthrough. Twenty years later, asked about it again, he was unembarrassed about the principle: “I’m not kind of arguing against the principle. I was very pro, actually, I thought there was a real clear case for it.”

Two decades of hindsight, and the case is still clear to him. What that case rests on is a claim about the relationship between a person and the state, and it is the same claim underneath every age check and identity gate now being built. The Identity Cards Act reached the statute book on 30 March 2006 and was repealed on 21 January 2011.

Then came 2008, and Culture Secretary Burnham produced a plan that looks like a blueprint for everything running today. The organizing idea was that online content should meet the same regulatory standards as television. He wanted taste and decency rules applied to the web, cinema-style age certificates and warning signs on anything featuring sex, violence or strong language, alerts bolted onto viral content, and ISP-level “child-safe” packages.

He floated making YouTube and Facebook delete offensive material within a set time once alerted, and changing libel law so suing an online publisher got easier. He even wanted to sound out the incoming Obama administration about international rules for English-language websites. Age gating, takedown clocks, regulator muscle aimed at platforms, it is all sitting there in 2008, waiting for the technology to catch up.

The plans died, and the body that killed them was Ofcom, which called them undesirable and unworkable online. This was back when the speech regulator was more levelheaded. The regulator that told Burnham in 2008 that broadcast rules could not be stretched over the internet is the regulator now enforcing the Online Safety Act. The intellectual weight at Ofcom has obviously severely declined.

Keep reading

HOPE Not Hate Deletes Page Listing Slain Brexiteer Ann Widdecombe’s ‘Dangerous’ Opinions

The radical leftist HOPE Not Hate (HnH) group has deleted a section of its website detailing the supposedly offensive opinions of former Brexit Party politician Anne Widdecombe, who was allegedly murdered in her country home earlier this month.

Following the killing of Ann Widdecombe, 78, a former Conservative government minister and Member of European Parliament for Nigel Farage’s Brexit Party, HOPE Not Hate removed a publication in which it accused the Christian politician of expressing “homophobic” and “sexist” opinions, while listing numerous supposed offences.

The HnH post stated: “Despite remaking her image in recent years through her appearances on reality TV shows, Widdecombe has a long history of homophobic and sexist remarks that jar with [Nigel] Farage’s claims that his party is ‘intolerant of intolerance’.

“Indeed, embracing the dangerous and divisive politics of her new vehicle, in the first Brexit Party speech in the European Parliament, Widdecombe made a risible comparison of Britain’s departure from the EU to the emancipation of slaves.”

In a statement to The Telegraph, a spokesman for the leftist group said that it was a “straightforward editorial decision” to remove the Widdecombe section of their website, and that it was done as a “mark of respect”.

“The article was a factual account of her political record, party affiliations and public statements, and we stand by its accuracy. Its removal reflects no change in our approach: scrutinising the public records of politicians is a legitimate and essential part of journalism and of our work, and it will continue.”

“Our thoughts are with Ann Widdecombe’s family and friends, and we share the widespread shock at her death,” the organisation added.

Despite initial insistence from police that there were no indications of a political motive, the alleged murderer of Widdecombe is now being investigated as a potential act of terror by Counter Terrorism Policing South East. It has also been reported that police sweeps of the current suspect’s home contained far-left communist material.

Keep reading