By Alina Maria Stan   The Trump administration announced on Wednesday that the Tesla Model Y is the first vehicle to pass NHTSA’s new advanced driver assistance safety tests. The same agency is simultaneously investigating 3.2 million Tesla vehicles for crashing while using the company’s more advanced self-driving system. The announcement celebrates Tesla for passing a test that measures […] This story continues at The Next Web

Source:: The Next Web

By Cristian Dina   Rocket Lab’s revenue grew 64 per cent, its stock hit a record high, and its backlog reached 2.2 billion dollars. The company sold more launches in the first quarter of 2026 than in the entire previous year. The only thing that has not launched yet is the rocket the market is pricing in. First-quarter […] This story continues at The Next Web

Source:: The Next Web

By Shikhar Mehrotra For the first time, students and educators can buy an Apple Watch through Apple’s Education Store, with the Series 11 down to $359, SE 3 to $229, and Ultra 3 to $719 in the US.

Source:: Digital Trends

Echoing concerns from other security experts, Orange Cyberdefense (OC) recently warned that employees have become the biggest security threat faced by business. 

Now, in the latest illustration of its ongoing security response, Apple is putting new protections in place in macOS 26.4 that should help – but employee education remains critical as hackers turn to complex, multi-stage, social engineering attacks to infest systems with malware.

Your people are your weakness

The data tells its own story. OC explains: Employees account for 57% of all security incidents and 45% of these incidents come when workers bypass or ignore security policies by, for example, using unapproved tools. 

Attackers are actively searching for and exploiting those kinds of policy workarounds, seeking weaknesses in commonly used, but unapproved, tools. Users really should educate themselves.

While companies can put some mitigations in place using device management and policy controls to constrain app use and downloads across their endpoints, Apple is also working to keep systems secure with a focus on the Terminal app. 

Terminal’s early warning system

In this case, it will introduce new malware warnings and protections to help prevent people from using Terminal to override system security to install malware-laden scripts. That’s the attack vector currently being used in the ClickFix series of attacks, which use fake macOS utilities to trick Mac users into doing just that.

It’s yet another example of how attackers rely on complex social engineering attacks to fool targets into undermining their own security. These attacks often begin with an attempt to get users to install infostealer malware on their own machines, and run them, bypassing Mac’s native malware defence.

Apple already has many, many protections to help combat attacks like these; now, we’ll see warnings in macOS Tahoe 26.4 whenever a relatively novice user pastes anything into the Terminal. Apple’s XProtect continues to block known malicious scripts. 

Helping people make better decisions

These warnings don’t appear in the first 24 hours after setting up a Mac, nor do they appear if a user has developer tools such as Xcode installed. That’s because Apple assumes developers are savvy enough to avoid falling for such tricks, while many users setting up their Macs may have legitimate need to use Terminal for legitimate purposes. (Apple will always warn when you try to paste code from sources known to be malicious.)

To an extent, Apple’s new protection reflects its belief that users should have choice while ensuring they are informed. Figuring out when to warn a user of the dangers they take has always been a challenge, as you don’t want to interfere in the user experience too heavily. But the prevalence of the kinds of threats OC warns about pushed Apple to put a new gate in place. 

FileVault keys come to the Passwords app

This isn’t the only new protection Apple has planned for macOS 26.4. The update does something many have long wanted. Ever since Apple’s first M-series chips arrived, we’ve had situations in which users forget their FileVault key, which can lead to Macs getting bricked when sold. Apple has now moved the macOS FileVault recovery key into users’ end-to-end encrypted Passwords app.

That’s good in two ways: it removes the threat Apple could lose or leak the key and makes it easier for a user to recover that key using the Passwords app on anther device. When you protect the data on your Mac with FileVault, you get a recovery key during set-up. If you forget the password for your Mac, you can reset the password by entering the recovery key.

Finally, IT admins seeking to ensure compliance with security policies will appreciate that Apple began rolling out Background Security Improvements in iOS 26.3.1, iPadOS 26.3.1 and macOS 26.3.1 to deliver incremental fixes and additional protections in between normal software updates. Still, as the OC data shows, the best and most effective security (beyond moving to a Mac) is to ensure employees fully understand the implications and significance of your company’s current security policies.

Please follow me on social media! Join me on BlueSky,  LinkedIn, and Mastodon.

Source:: Computer World

By Vikhyaat Vivek Meta is reportedly building AI tools that can complete everyday tasks, including an Instagram shopping assistant planned for late 2026.

Source:: Digital Trends

By Vikhyaat Vivek South Korea’s largest Buddhist order has introduced Gabi, a humanoid robot monk, raising fresh questions about how AI and robotics may fit into spiritual life.

Source:: Digital Trends

By Deepti Pathak Google has announced a major upgrade to its fitness and wellness platform by introducing Fitbit Air…
The post Google Just Killed Whoop With Fitbit Air & New Health Coach appeared first on Fossbytes.

Source:: Fossbytes

By Deepti Pathak HMD plans to introduce the Vibe 2 5G in India on May 8. Ahead of the…
The post HMD Vibe 2 India Launch & Flipkart Sale Confirmed appeared first on Fossbytes.

Source:: Fossbytes

AI is capable of mimicking a real person. It’s clear this capability exists, and the ethics of using AI for this purpose are often very clear. But increasingly, new applications are leading to ethically murky results. 

The good

For example, the CEO of a company, or a politician, could choose to create a clone using AI tools, creating a chatbot plus an avatar — a digital twin — that can interact with people on their behalf. Silicon Valley is big on the idea: Meta’s Mark Zuckerberg and LinkedIn co-founder Reid Hoffman are working on, or have already created, digital twins of themselves. 

Cloned politicians include Pakistan’s Imran Khan, who used an authorized voice clone to campaign from prison, and New York City Mayor Eric Adams, who used voice-cloned robocalls to speak with constituents in languages like Mandarin and Yiddish.

This kind of use case is probably ethical — as long as the people interacting know that they’re dealing with a digital clone and not a real person. 

The bad

The flip side of ethical uses for AI-generated clones is the non-consensual (and therefore unethical) cases. And of these, there are already many. For instance:

In 2019, the first widely documented case occurred when scammers used AI to mimic the voice and German accent of a parent company’s executive, successfully tricking the CEO of a UK energy firm into transferring €220,000 into a fraudulent account.

In 2023, an Arizona mother, Jennifer DeStefano, was targeted by extortionists who used an AI clone of her 15-year-old daughter’s voice to demand a $1 million ransom.

And in 2024, a finance worker at a multinational firm in Hong Kong was tricked into transferring $25 million after attending a video conference call featuring deepfake recreations of his CFO and several other colleagues. 

Other unethical, non-consensual uses for AI cloning include deepfake videos, where a celebrity’s face is superimposed on a porn actor. In all the above examples, the ethics are clear. This is all very wrong. 

But with China leading the way in the emergence of AI clones, the ethics are becoming far murkier. 

And the ugly

One emerging trend involves workers using specialized software to build digital versions of their bosses or colleagues. The most prominent project driving this trend is Colleague Skill, which was posted in late March by its creator, a 24-year-old Shanghai-based engineer named Zhou Tianyi. 

Colleague Skill and its forks and copycats, which tend to be open source, enable people to upload chat histories, emails, and internal documents to create a functional persona that mimics a specific coworker’s professional expertise and communication style. The technology stack includes tools like Claude, Kimi, ChatGPT, DeepSeek API, OCR (Tesseract), and sentiment analysis modules.

Colleague Skill uses a person’s past communications to build a talking replica of their personality. If you think of a regular AI as a general student who knows a little bit about everything, this tool acts like a specialized mask that forces the AI to behave like one specific individual. 

In other words, it produces a chatbot with the knowledge and patterns of speech of a real person. 

Colleague Skill started as a satirical commentary on AI-driven layoffs. But some employees began using it in earnest to clone their colleagues. There are several stated reasons for doing so, including retaining institutional knowledge and having an instant sounding board to “discuss” plans and ideas with. 

A similar motivation is the use of AI to clone bosses, so employees can better predict how that boss might react to the employees’ work. 

In most of these instances, according to reports out of China, the creation of the boss-bot or colleague clone is nonconsensual. 

Is non-consensually basing a custom chatbot on a colleague or boss unethical? 

And then it got personal (and weird)

Tianyi, creator of Colleague Skill, later forked it into something called Ex-Partner Skill. The idea is to re-create a former partner with AI so the user can continue the relationship. 

It operates on the same technical engine but applies it to a much more personal part of life. Users upload photos, social posts, chat logs and other content. The AI chatbot can then mimic the former partner’s tone, catchphrases, and subtle linguistic nuances, something that, “truly sounds like them — speaks with their catchphrases, replies in their style, remembers the places you went together.”

This allows a person to simulate conversations with someone who is no longer in their life.

If Colleague Skill is in a grey area, Ex-Partner Skill is in a darker grey area. 

(Note: many of the original repositories for Ex-Partner Skill have been removed from public view in China or “sanitized” after regulatory pressure. But the framework reportedly continues to circulate in private developer circles, and similar tools are increasingly used for “digital resurrection.”)

Ethically, the concept feels like it exists on a wide spectrum somewhere between therapy at one end and revenge porn at the other. (It’s like revenge porn in the sense that when “content” consensually made by two people for one purpose is later used consensually by one person in a way that the other person might find objectionable.)

Or maybe it’s closer to the “deathbot” phenomenon, where an AI-generated simulation provides a fake version of the dearly departed. (In both cases, the user interacts with a digital twin of someone who is no longer present in one’s life.) In fact, some people in China are using Ex-Partner Skill as a deathbot for a deceased loved one. 

The lack of consent feels like an ethical lapse. But we don’t consider it unethical to think about, remember, imagine conversations with, or journal about ex-partners — or dead family members. 

Boosters of the Ex-Partner Skill idea say that conversations with digital exes are therapeutic. They point out that because it’s private, it’s not harassment or stalking or an invasion of privacy. Instead, they argue, it helps with personal reflection and emotional healing.

As for people who have died, according to Chinese media reports, some users say the tool gives them a sense of closure and allows them to say the things they wish they could have said to the real person. But is it really closure if one person is still obsessively trying to interact — or pretend to interact — with the other person?

It’s healthy to communicate. But it’s not communication when a person is by themselves talking to no one and sending messages to a person who never gets those messages.

While ex-bots are a thing these days in China, the trend is showing up elsewhere. Some Character.AI users outside of China have created chatbots based on ex-partners, even though  the company has changed its Terms of Service to explicitly ban the creation of bots using the likenesses of private individuals without their permission. 

The emergence of nonconsensual cloning of coworkers, bosses and ex-partners is a new challenge to our sense of right and wrong, and yet another way AI is challenging us to step up and figure out how to respond.

Source:: Computer World

By Sudhanshu Kumar Mangalam AirPods with cameras sound wild, but they might be exactly what Siri needs to stop feeling stuck in the past.

Source:: Digital Trends

By Alina Maria Stan   The person who connected Jeff Bezos to one of the most ambitious electric vehicle startups in America has left its board. Melinda Lewison, who manages the Bezos family office and was listed as a director on Slate Auto’s corporate filings, has departed the company’s board months before its first truck is scheduled to roll […] This story continues at The Next Web

Source:: The Next Web

A LinkedIn feature that allows paid subscribers to view a list of visitors to their profile should be made available to all EU users free of charge to comply with the region’s General Data Protection Regulation (GDPR), a legal complaint launched by the None of Your Business (NOYB) digital rights group has claimed.

Filed this week in an Austrian court, the group’s argument is that LinkedIn’s ‘Who’s Viewed Your Profile’ feature contravenes the GDPR Article 15, which covers a subject’s right of access to their own data.

NOYB has a history of taking on tech companies. In 2025, Google was hit by a €325 million ($381 million) fine by French privacy regulator, the CNIL, over its data collection and advertising policies after a complaint by the group.

Contradictory policy

LinkedIn began offering users the ability to see who has viewed their profile around 2007, later turning this into a paywalled perk in a move that pre-dated the arrival of GDPR in 2018.

According to NOYB, this commercialization left non-subscription users in a bind. Profile visitor data should legally be accessible to EU citizens under GDPR, but when they ask for this via a formal Data Subject Access Request (DSAR), LinkedIn refuses access, citing data protection.

Despite this, if the user subscribes to a LinkedIn Premium Career plan starting at €30 per month ($40 per month in the US), the same data suddenly becomes accessible.

“It is particularly absurd that LinkedIn is using a supposed ‘data protection interest’ as an argument to deny the right of access to data under the GDPR,” argued NOYB’s press release.

In NOYB’s view, LinkedIn’s policy is contradictory. The company limits access to something that should legally be free because allowing access would undermine the incentive to pay for it.

“Either the data must not be accessible to anyone, or – if it is clear to the visitor that the data is visible – it must also be disclosed in accordance with Article 15 GDPR,” NOYB said. In its view, LinkedIn’s policy of charging to access this data is illegal and the company should be fined to prevent future breaches.

Right to view

LinkedIn will doubtless point out to the Austrian Data Protection Authority that all users, including free subscribers, can opt out of having their profile visit made visible by toggling off the feature in Settings/Visibility tab/’Visibility when viewing other profiles’. Then each visit a user makes to another profile is recorded as one by an ‘Anonymous LinkedIn Member’. Free users can also see the last five visitors to their profile, as long as those users have not selected this anonymity setting.

It’s possible the company will further argue that, under Article 15, the rights of users to know who has viewed their data conflicts with the rights of other users to maintain their own privacy.

When contacted for response, a LinkedIn spokesperson sent the following statement: “This assertion [by NOYB] is false. Not only is it incorrect that only Premium members can see who has viewed their profile, but we also satisfy GDPR Article 15 by disclosing the information at issue via our Privacy Policy.”

According to Helen Brain, partner and head of commercial at Square One Law in the UK, the case would cause problems for LinkedIn’s lawyers even if the outcome remained uncertain.

“NOYB appears to have a strong argument that LinkedIn is breaching GDPR in one way or the other, but it’s impossible to say how likely they are to succeed before we see LinkedIn’s counter-arguments,” she said.

The complaint is on strong ground when arguing that profile visits should fall under GDPR Article 15 Right of Access. “If the viewer’s personal data is private and shouldn’t be disclosed in response to a DSAR by the viewed person, logically that means the viewer’s personal data should not be disclosed to premium account holders either,” said Brain. “If NOYB is successful in its complaint, the Austrian Data Protection Authority could ultimately issue a fine, and that could be substantial.” 

However, predicting the wider effect on technology companies using the same ‘data as a feature’ to incentivize paid subscriptions is difficult in advance of a ruling. If NOYB prevails, LinkedIn could be ordered to stop its disclosure of profile searchers or, alternatively, to make this available free of charge in response to DSARs.

However, Brain believed the issue might come down to the way consent is gained. “Even if LinkedIn is ordered to change what it is doing, it will find a new way to gain consent to permit the disclosures of searchers lawfully and continue to charge for the data they gather.”

Source:: Computer World

By Sudhanshu Kumar Mangalam Snapchat has walked away from its Perplexity AI search plan, raising fresh questions about how much AI users actually want inside social apps.

Source:: Digital Trends

By Alina Maria Stan   Tesla has filed a trademark for a bespoke Roadster badge that looks like it belongs on a Lamborghini. The car it will adorn was first promised nine years ago. A prototype debuted in November 2017 with a 200 kilowatt-hour battery, a claimed 620-mile range, a 1.9-second zero-to-60 time, and a starting price of 200,000 […] This story continues at The Next Web

Source:: The Next Web

By Shikhar Mehrotra The Fitbit Air weighs 5 grams without its strap, tracks heart rate, blood oxygen, skin temperature, and steps around the clock, lasts seven days on a charge, and costs $99.99.

Source:: Digital Trends

European Union member states and the European Parliament agreed early Thursday to push back the toughest deadlines under the bloc’s AI Act, giving enterprises more time to prepare for high-risk compliance.

Under the provisional deal between negotiators for the European Parliament and European Council, high-risk AI systems will face new deadlines of Dec. 2, 2027 for stand-alone systems and Aug. 2, 2028 for AI used in products covered by EU sectoral safety rules, a European Parliament statement said. The original deadline was Aug. 2, 2026.

The deal still needs formal adoption by both Parliament and Council before it can enter into law. The co-legislators intend to complete that step before Aug. 2. Until they do, the original deadline applies as drafted.

“Today’s agreement on the AI Act significantly supports our companies by reducing recurring administrative costs,” Marilena Raouna, Cyprus’s deputy minister for European affairs, said in a statement from the Council, which is composed of representatives of each of the EU’s 27 member states. Cyprus holds the rotating presidency of the Council, which negotiates on behalf of member states.

The breakthrough comes nine days after previous discussions collapsed without agreement.

Fewer restrictions, more time to implement

The provisional agreement removes overlapping rules for AI in machinery products, Parliament said. These will now follow only sectoral safety rules, with safeguards meant to ensure equivalent health and safety protection.

It also narrows what counts as a “safety component” under the AI Act. AI features that only assist users or improve performance will not automatically be treated as high-risk, the Parliament said, as long as a failure does not create health or safety risks.

For wider sectors such as medical devices, toys, lifts, machinery and watercraft, the co-legislators agreed on a mechanism to resolve overlaps between the AI Act and existing sectoral laws, the Council said in its statement.

The deadline for member states to set up AI regulatory sandboxes has been pushed back by a year to Aug. 2, 2027, the Council said. Watermarking obligations on AI-generated content, on the other hand, will apply earlier than the Commission proposed, from Dec. 2, 2026 instead of Feb. 2, 2027, the Parliament said.

Mid-size firms get more breathing room. Exemptions previously available only to small and medium-sized enterprises now extend to small mid-cap companies, the Council said. The deal also clarifies that the EU’s AI Office will supervise general-purpose AI systems centrally, with national authorities keeping responsibility in areas including law enforcement, border management, judicial authorities and financial institutions.

“With this agreement, we show that politics can move just as quickly as technology,” said Arba Kokalari, the Parliament’s co-rapporteur for the Internal Market and Consumer Protection committee. “We now make the AI rules more workable in practice, remove overlaps and pause the high-risk requirements.”

Parliament and Council also agreed to ban AI systems that create child sexual abuse material or that depict identifiable people in sexually explicit content without consent, the Parliament said. The ban covers placing such systems on the EU market, doing so without safety measures to prevent misuse, and using them to generate the content. Companies have until Dec. 2, 2026 to comply.

“Alongside simplification measures, we are banning nudification apps, a key part of the Parliament’s mandate, and, of course, the creation of child sexual abuse material using AI systems,” said Michael McNamara, the Parliament’s co-rapporteur for the Civil Liberties, Justice and Home Affairs committee.

What still applies

Several parts of the AI Act keep moving on their original schedule. Bans on unacceptable-risk AI have applied since February 2025, according to the European Commission. The general-purpose AI rules came into force in August 2025. The transparency obligations under Article 50, including disclosure for chatbot interactions, are set to apply from Aug. 2, 2026.

The provisional agreement is part of the seventh omnibus package on simplification, proposed by the Commission on Nov. 19 last year in response to the Draghi report on EU competitiveness.

Source:: Computer World

By Deepti Pathak The vivo X300 Pro was perhaps the best phone of 2025, and still is. But vivo…
The post Vivo X300 Ultra and X300 FE Debut in India: Price, Features and More appeared first on Fossbytes.

Source:: Fossbytes

By Rachit Agarwal A new study from researchers at MIT, Carnegie Mellon, Oxford, and UCLA found that even brief AI use can hurt your ability to problem-solve once the tool is taken away.

Source:: Digital Trends

By Alina Maria Stan Google has a publisher problem. AI Overviews, the AI-generated summaries that now appear at the top of search results for a growing share of queries, have been correlated with a 58 per cent reduction in click-through rates to the websites whose content those summaries are built on. Penske Media has filed an antitrust lawsuit. The […] This story continues at The Next Web

Source:: The Next Web

By Alina Maria Stan Snap reported first-quarter earnings on Tuesday that should have been unremarkable: revenue up 12 per cent to 1.53 billion dollars, adjusted EBITDA more than doubled to 233 million dollars, free cash flow nearly tripled to 286 million dollars. The stock fell four per cent. The reason was not in the numbers Snap reported but in […] This story continues at The Next Web

Source:: The Next Web