AI was meant to cut costs. Why is it becoming the biggest new expense?

by Hansraj Agrawal | Jul 2, 2026 | Technology

The adoption of artificial intelligence (AI) has grown sharply over the past couple of years. Much of this rise stems from AI being positioned as the ultimate productivity tool— one that can automate repetitive tasks, accelerate decision-making, and help businesses do more with less. For companies, the promise lies in lower operating costs and greater efficiency. For professionals and freelancers, AI offers the possibility of replacing routine administrative work with a round-the-clock digital assistant.

However, AI is no longer just a technology investment; it is becoming a recurring business expense. Monthly subscriptions, enterprise licences, usage-based pricing and investments in employee training are creating a new line item in corporate budgets. The conversation, therefore, is beginning to shift—from “should we adopt AI?” to “are we getting enough value from what we spend?”

‘Artificial’ intelligence: The real cost

Recent incidents have brought AI spending into focus. Media reports have flooded the internet with a configuration oversight that allegedly led to an astronomical bill for Anthropic’s Claude AI, which served as a reminder of how usage-based pricing can escalate unexpectedly. Around the same time, reports suggested that Amazon had asked its employees not to use AI merely for the sake of increasing usage after shutting down an internal AI-consumption leaderboard.

While the two incidents are unrelated, they show the latest trend that companies are now becoming more conscious not just of AI adoption, but also of AI economics.

Unlike traditional software that was often purchased through annual licences, many generative AI services operate on subscription models or charge based on token consumption. As employees increasingly rely on AI for writing, coding, research, customer support and data analysis, these costs can grow rapidly.

Looking beyond expense

Rising AI costs are also reshaping procurement decisions. Reuters recently reported that a growing number of companies are moving away from relying exclusively on premium AI models and are instead opting for smaller, less expensive alternatives for routine tasks, as usage-based pricing makes AI spending less predictable.

However, experts caution against viewing AI purely through the lens of expense.

“The ‘AI Cost’ is to be looked at as an investment, where organisations are readying a blueprint for the future, where AI intelligence and human intelligence can co-exist,” said Raghu A, Partner, Deloitte India, in response to queries from Business Standard.

According to him, organisations are embedding AI into their long-term strategy rather than deploying it merely as a productivity tool. “We have to open ourselves to the possibility of AI being more than just an efficiency or productivity solution,” he said, arguing that businesses should view AI as a value creator rather than simply a means to improve the top or bottom line.

The ROI question

Even as AI becomes more deeply embedded in business operations, executives are under pressure to demonstrate measurable returns.

Evaluating AI outcomes begins much before measuring financial returns, Raghu said. He added that companies first need a structured adoption framework, identify business use cases, equip employees with AI skills, develop applications and drive adoption across the organisation.

Once these foundations are in place, organisations should measure AI’s impact across four dimensions: direct profit and loss impact, indirect business value, efficiency gains and opportunity creation.

Importantly, Raghu noted, “AI investments also take time to mature, meaning value realisation cannot always be expected immediately.”

That reflects a broader shift in enterprise AI adoption. The initial race to deploy AI tools is gradually giving way to a greater emphasis on governance, accountability and measurable business outcomes.

Another monthly bill?

The economics of AI are also changing for smaller businesses and independent professionals.

Many freelancers today pay for premium AI subscriptions to assist with research, writing, coding or design tasks that might otherwise require hiring additional help. For some, AI has become as essential as cloud storage, accounting software or high-speed internet.

Whether AI ultimately replaces human assistance or simply becomes another unavoidable subscription remains an open question. Raghu believes the technology should be viewed in terms of the value it creates. “AI is a tool that SMBs and freelancers should use effectively to bring about the difference between expense and value,” he said.

ALSO READ: OpenAI discusses giving US govt 5% stake as Trump weighs public share in AI

Not always useful

AI may be a powerful tool, but it cannot always deliver results without human expertise. That was recently evident at Ford Motor Co., which rehired around 350 veteran engineers over the past three years after finding that AI failed to match their skills and replace decades of experience in vehicle engineering and quality control.

The company brought back former employees and experienced supplier engineers to mentor younger staff, review designs, identify manufacturing issues and help train AI systems. Rather than replacing either humans or AI, Ford is using the two together. According to the BBC, company executives said the “technology is only as good as the knowledge and data behind it.”

What happens 5 years from now?

If the past two years were about AI adoption, the next phase is likely to be about AI optimisation. Companies are expected to scrutinise subscriptions, monitor usage more closely and prioritise cases that deliver measurable returns.

Five years from now, AI may well become as indispensable to business operations as the internet or cloud computing. But unlike those technologies, its value may not be judged by how widely it is used, but by whether it creates outcomes that justify its growing place on the balance sheet.

Source link

WhatsApp usernames rollout under scanner: What the govt is worried about

by Hansraj Agrawal | Jul 2, 2026 | Technology

The Ministry of Electronics and Information Technology, and the Ministry of Home Affairs is set to meet representatives of several peer-to-peer messaging platforms today (July 2) to discuss the implications of usernames on WhatsApp and potentially similar features elsewhere. The meeting follows a notice Meity sent WhatsApp asking it to defer the rollout of its new usernames feature, which lets users message each other without sharing phone numbers, until the ministry is satisfied the feature cannot be exploited for fraud. WhatsApp has pushed back, saying it has built multiple safeguards against impersonation and scams. Here is what has happened so far.

What WhatsApp announced

WhatsApp began rolling out username reservations on June 29, a feature that lets users connect with others without sharing their phone numbers. The system itself is not live yet, but users can reserve a unique username ahead of rollout planned later this year. Once live, users will be able to share a username instead of a phone number when starting a new conversation, and WhatsApp has said there will be no public directory or username suggestions, so someone will need to know the exact username to reach another user for the first time. An optional username key adds a further layer, requiring anyone messaging a user for the first time through their username to enter a code before the conversation can begin.

A phone number will still be mandatory to create and use a WhatsApp account. The username does not replace the number for registration, it offers an alternative way to connect with new contacts without revealing that number during the first interaction, and users can change or remove their username whenever they want. Businesses and creators will also be able to claim their existing Instagram or Facebook usernames on WhatsApp to keep their identity consistent across Meta’s platforms.

What are the concerns

Meity has asked Meta to defer the rollout until consultations are completed, citing concerns over impersonation, fraud and online scams, and has directed the company to submit a detailed explanation within three days. Business Standard reported that the Home Affairs Ministry first raised concerns about cyber fraud and conveyed them to Meity, prompting Thursday’s meeting with executives from several messaging platforms. Business Standard also reported that the examination would focus on whether the feature could be exploited by scammers, and that concerns about genuine users being unable to claim their own names were also being flagged to WhatsApp.

PTI reported that the government, as the statutory authority, needs to satisfy itself that the feature does not pose risks before it can proceed, and that if WhatsApp’s response is unsatisfactory, it may seek to prevent the rollout altogether.

Much of this concern stems from how usernames have played out on other platforms such as Telegram and Signal.

As usernames are self-chosen and easy to alter, threat actors can craft handles that closely mimic bank executives, officials or corporate brands, making phishing attempts look far more credible than a random phone number would.

Public usernames within large groups also make it easier for bad actors to scrape identities and build profiles of potential targets, feeding into more targeted scams and open source intelligence gathering.

There is also the problem of account recycling, since a username that is changed or deleted goes back into the public pool, letting someone else claim a recognisable handle and deceive contacts who assume they are still messaging the original user.

What WhatsApp has said

WhatsApp has pushed back on the idea that the feature is unsafe, saying it has built layered defences against impersonation and has safeguards in place such as limitations on how many new people an account can contact. In a statement, a WhatsApp spokesperson said:

“We’ve announced the option for people to reserve their preferred username on WhatsApp. The ability to use a username is not yet live and will roll out slowly later this year. To protect against impersonation, we’ve held the highest-profile names — think public figures, government entities, celebrities, verified Meta accounts — so they can only ever be claimed by their legitimate owners and lookalike derivatives of known names are held as well,” said WhatsApp on the impersonation fears raised by many over usernames usage.

Additionally, it said that users will still require a phone number to use WhatsApp and it has built multiple layers of defense against scams into usernames. These include: Other users need to know the exact username to message you, we will limit how many new people an account can contact, block repeated attempts to guess someone’s username key, and have systems to detect and remove activity showing common impersonation and abuse patterns.

“When the feature becomes available and someone sends you a message for the first time via your username, we will show you if they’re a new account, if they’re your contact, if you have groups in common, and if they’re based in a different country, so you can decide whether to respond,” said WhatsApp.

WhatsApp has also posted a detailed FAQ thread on X that addresses most of Meity’s concerns directly, covering how usernames can be reserved, what happens when someone tries to impersonate a well-known name, and how the username key limits unwanted contact.

It also noted that usernames, like phone numbers, cannot be searched within the app, and that the strongest safeguard against unwanted contact remains enabling a username key, which can be reset at any time to cut off new inbound messages. It also said that Linking an Instagram or Facebook account is only required if a user wants to match that handle’s username on WhatsApp, and the accounts can be unlinked afterward.

What legal advocates say

The Internet Freedom Foundation, a digital rights advocacy group, has pointed out that no clear legal provision forms the basis of Meity’s notice to Meta, in a post on X. It stated that Section 79 of the IT Act is a safe harbour provision governing when a platform can be held liable for user content, which it says does not necessarily extend to approving product features before release. It also referred to Sections 66C and 66D, which punish identity theft and cheating by personation, stating that these provisions are aimed at individuals who misuse a tool rather than at the company that builds it.

On the IT Rules, IFF stated that Rule 3(1)(b), Rule 3(2) and Rule 4 set out due diligence and grievance obligations, which it said do not necessarily translate into a licensing mechanism for approving features. It also referred to Section 69A, the provision that allows Meity to block specific online content through a defined procedure, stating that this does not necessarily cover vetting features before launch.

The group also referred to a March 2024 instance, when Meity asked large intermediaries including AI companies to seek permission before deploying under-tested AI models, a requirement it withdrew within a fortnight after it was flagged as lacking a clear empowering provision.

Will this affect businesses

Business Standard previously reported that executives across the direct to consumer and logistics ecosystem see limited impact on how they engage with customers, largely because phone numbers remain central to commerce and delivery. Here is what the executives from the direct-to-consumer (D2C) and logistics ecosystem told Business Standard.

Source link

Google loses EU top court battle over $4.7 bn Android antitrust fine

by Hansraj Agrawal | Jul 2, 2026 | Technology

By Peter Chapman and Samuel Stolton

Google lost its long-running fight against a €4.1 billion ($4.7 billion) European Union antitrust fine after the bloc’s top judges said regulators were right to punish the US giant for abusing Android’s market power.

The European Court of Justice ruled on Thursday that Google’s earlier defeat against a European Commission penalty should stand. The decision is legally binding and marks a significant win for the Brussels-based regulator, which has been fighting Google through the EU’s courts since the fine was first leveled in 2018.

“The appeal brought by Google and its parent company Alphabet against the judgment of the General Court is dismissed, thereby confirming the penalty imposed for Google Search’s abuse of a dominant position in the context of the Android operating system,” the court said in a statement on the judgment.

The decision is a constraint on the Android business model — which has provided free software in exchange for conditions imposed on mobile phone manufacturers. Such contracts provoked the ire of the commission in 2018, when the watchdog accused Alphabet Inc.’s Google of three separate types of illegal behavior that helped cement the dominance of its search engine, accompanying the order with the then-record fine. The decision also paves the way for a wave of potential lawsuits from victims of Google’s behavior.

Google said the ruling “fails to recognize our significant investment to ensure Android remains open, interoperable and free. In any event, we adapted our agreements to comply with the initial decision back in 2018 and we remain focused on continued innovation and openness for our users, partners and developers.”

FairSearch, a group of complainants that brought the case to the commission in 2013, called the ruling “an important victory in Europe’s highest court against Google’s anti-competitive conduct in mobile markets.”

Play Store

In its decision to fine Google, the commission said it was illegally forcing handset makers to pre-install the Google Search app and the Chrome browser as a condition for licensing its Play Store — the marketplace for Android apps.

Secondly, the EU said Google made payments to some large manufacturers and operators on condition that they exclusively pre-installed the Google Search app.

Lastly, the EU said the Mountain View, California-based company prevented manufacturers wishing to pre-install apps from running alternative versions of Android not approved by Google.

In a September 2022 ruling at the EU’s lower General Court, judges upheld the vast majority of the commission’s arguments, but cut the fine from €4.3 billion after finding that regulators hadn’t provided enough evidence for specific abuses.

The Android case was one of four against Google that formed the centerpiece of erstwhile EU competition chief Margrethe Vestager’s efforts to crack down on the growing power of big tech companies.

Since being replaced by Spanish official Teresa Ribera in 2024, the company has continued to garner EU antitrust scrutiny. Under the bloc’s powerful Digital Markets Act, Google was earlier this year told to lift technical barriers to rival AI search assistants on Android and give key data to other search engine providers.

Separately, the company faces upcoming penalties under the DMA over allegations it unfairly favors in-house services across its sprawling search empire and for preventing app developers from steering consumers to offers outside of its Play Store, and is also being probed over concerns it unfairly demotes certain news results.

Source link

SC tears into AI-hallucinated judgments, sets aside NCLT insolvency order

by Hansraj Agrawal | Jul 2, 2026 | Technology

The Supreme Court on Thursday set aside an order of the National Company Law Tribunal (NCLT) in the insolvency case of Essel Infraprojects after finding that the tribunal had relied on non-existent, AI-generated judgments. The fake citations were made in the case of Pooja Ramesh Singh vs Jammu and Kashmir Bank Ltd.

Why did the Supreme Court set aside the NCLT order?

A Bench of Justices P S Narasimha and Alok Aradhe held that reliance on hallucinated legal material “strikes at the integrity of adjudication and its processes” and directed courts to adopt a “zero-tolerance” approach towards the production or use of AI-generated precedents without proper verification.

“Courts must adopt a zero-tolerance mode for producing, citing or using AI-generated precedents without verification,” the Bench said.

The Court held that citing fake AI-generated precedents amounts to “professional misconduct” by advocates. It also observed that it is a serious judicial lapse if a judge relies on such fabricated material while deciding a case.

“We have no hesitation in declaring that such a decision is no decision in the eyes of the law,” the Court said, adding that a judgment must be set aside even if “an iota” of fake or hallucinated material enters the decision-making process.

Accordingly, the apex court set aside both the NCLT order and the National Company Law Appellate Tribunal (NCLAT) judgment that had upheld it, directing the tribunals to reconsider the matter solely on the basis of verified facts and applicable law.

What did the Supreme Court say about AI in adjudication?

The Bench also raised concerns over the use of artificial intelligence (AI) in the judicial process. While AI can be used as an assistive tool, “human control over adjudication must remain total and absolute at every stage”, it said.

The Court acknowledged that professionals may increasingly rely on AI to manage growing workloads, but cautioned that unregulated use could erode independent reasoning and make legal professionals overly dependent on technology.

Unlike other technologies already integrated into courts, AI is capable of influencing thinking, reasoning and decision-making, the Bench observed, warning judges to exercise caution while using such tools.

The Supreme Court also directed the Bar Council of India to constitute a committee to examine the issue in detail.

What was the Essel Infraprojects insolvency case about?

The case arose from a plea filed by suspended director Pooja Ramesh Singh challenging the NCLT’s August 28, 2024, order admitting Essel Infraprojects into the corporate insolvency resolution process on an application filed by Jammu and Kashmir Bank over an alleged default of ₹87.43 crore. The NCLAT upheld the order on September 11, 2025.

While examining the appeal, the Supreme Court found that the NCLT had relied on non-existent, AI-generated precedents to support its reasoning and consequently set aside both tribunal orders for fresh consideration.

Source link

WhatsApp issues FAQs on username feature amid government scrutiny

by Hansraj Agrawal | Jul 2, 2026 | Technology

Meta-owned WhatsApp has come out with a detailed set of frequently asked questions (FAQs) on the controversial username feature, outlining details on how it plans to address concerns around impersonation, scams and unwanted contact as users begin reserving usernames.

The FAQs come amid the Centre’s notice to Meta flagging concerns over potential fraud and impersonation risks, and warning against a roll out of the feature till consultations on the issue are completed “to the satisfaction of the government”.

Put simply, the new feature will allow users to create unique usernames that can be used for connecting on WhatsApp without sharing phone numbers.

The FAQs say usernames will be optional, cannot be searched by strangers, and that users can add an additional ‘username key’ requiring both the username and the key before someone can contact them.

WhatsApp said well-known names, including those of public figures, celebrities, government entities and Meta-verified accounts, have been reserved so they can only be claimed by legitimate owners. It also said users linking their Instagram or Facebook accounts can claim matching usernames to help verify ownership, while retaining the option to unlink those accounts later.

The messaging platform said it will monitor ‘blocks’ and ‘reports’ to act against scammers, while cautioning that claims about reserving popular or well-known usernames are false. Only legitimate account owners can claim protected names, it asserts.

Here is a look at the key clarifications and safeguards outlined by WhatsApp in its FAQs: Are usernames mandatory? Nope, they are optional.

What if the username I want isn’t available? There’s a few reasons you might not be able to reserve the username you want: 1) It’s an existing Instagram or Facebook username; these are reserved for their owners.

2) We’ve held well-known names and some variations of them — like public figures, celebrities, government entities and Meta-verified accounts – so they can only be claimed by their legitimate owners. If you try to reserve those, the system will say it’s not available.

3) Someone already claimed a common name, in that case use the username generator.

What if someone creates a username similar to mine to impersonate me or run a scam, how do you stop someone from pretending to be me? Usernames are not available for messaging yet. When they are, and you get a message from someone new, we’ll let you know the country origin, and a warning for first time outreach. Well-known public-figure names and their variations are held for verified owners. We’re also keeping a close eye on blocks and reports to take action against scammers.

Can random people message me if they know or guess my username? Just like you can’t search for a phone number in WhatsApp, you can’t search for a username. The best way to prevent someone from contacting you is to add a username key and to choose a username that is unique to WhatsApp.

All the current measures remain in place to prevent unwanted contact, including warnings with details about unknown senders (whether they’re a new account, if you share groups, what country they’re in) and the ability to block and report.

What is a username key? An extra layer of protection you can enable with your username, so that another user will have to know both your username and your username key to contact you. You can reset your key at any time to stop new inbound contact through your username.

Do I have to link my other Meta accounts? If you want the same username as your Instagram or Facebook account, you’ll have to link them. This is one of the ways to reduce impersonation and to make sure you’re the legitimate owner of the account. But after that, you can unlink the account if you want, or you can choose a username unique to WhatsApp.

Can I change my username later?
Yes, as long as the new one you want is available.

A few more things to keep in mind…people are making false claims about reserving popular or well-known usernames – this isn’t true, only the legitimate account owners are able to reserve well-known public-figure names.

We enabled reservations before usernames launch later this year because we think people will feel strongly about what username they want on WhatsApp. We’re taking our time and listening to feedback so that when it rolls out later this year we get it right.

Source link

Serial entrepreneur Turakhia launches AI work platform Neo with $30 mn

by Hansraj Agrawal | Jul 2, 2026 | Technology

Serial entrepreneur Bhavin Turakhia on Wednesday announced the launch of Neo, an AI-native work platform designed to bring work, knowledge and AI execution into a single connected suite. He has committed $30 million of his own capital to build the company.

Turakhia’s fifth entrepreneurial venture enters the market at a time when enterprises are accelerating artificial intelligence (AI) adoption but continue to struggle to translate it into meaningful productivity gains. According to him, the challenge lies not in the quality of AI models but in the fragmented nature of enterprise work.

“Most organisations fail to capture the value of AI because context is fragmented, knowledge is scattered across teams and tools remain disconnected. Neo changes that by centralising context and making AI a first-class participant in every workflow, not a tab beside it,” Turakhia, founder of Neo, said.

The Neo Suite includes Friday, an AI assistant, co-worker and agent layer connected to more than 1,000 external applications; Tasket, a project management platform; Studio, a knowledge management platform for documents, spreadsheets and diagrams; and Drive, a file-sharing workspace where people and AI agents work together on files.

The launch comes as enterprises worldwide seek practical ways to deploy AI beyond experimentation and into core business operations. Neo is built on the premise that the next phase of enterprise AI will be defined by execution, with AI embedded directly into everyday workflows.

Turakhia has spent more than two decades building global technology companies. He previously co-founded Directi, Radix, Titan and Zeta. Directi’s web businesses were acquired for $160 million in 2014, Titan reached a valuation of $300 million following an investment from Automattic, while SoftBank-backed Zeta is currently valued at about $2 billion.

Source link

« Older Entries

Next Entries »

Times Top News

Times Top News