Tech
OpenAI has slipped shopping into ChatGPT users’ chats—here’s why that matters
Your phone buzzes at 6 a.m. It’s ChatGPT: “I see you’re traveling to New York this week. Based on your preferences, I’ve found three restaurants near your hotel. Would you like me to make a reservation?”
You didn’t ask for this. The AI simply knew your plans from scanning your calendar and email and decided to help. Later, you mention to the chatbot needing flowers for your wife’s birthday. Within seconds, beautiful arrangements appear in the chat. You tap one: “Buy now.” Done. The flowers are ordered.
This isn’t science fiction. On Sept. 29, 2025, OpenAI and payment processor Stripe launched the Agentic Commerce Protocol. This technology lets you buy things instantly from Etsy within ChatGPT conversations. ChatGPT users are scheduled to gain access to over 1 million other Shopify merchants, from major household brand names to small shops as well.
As marketing researchers who study how AI affects consumer behavior, we believe we’re seeing the beginning of the biggest shift in how people shop since smartphones arrived. Most people have no idea it’s happening.
From searching to being served
For three decades, the internet has worked the same way: You want something, you Google it, you compare options, you decide, you buy. You’re in control.
That era is ending.
AI shopping assistants are evolving through three phases. First came “on-demand AI.” You ask ChatGPT a question, it answers. That’s where most people are today.
Now we’re entering “ambient AI,” where AI suggests things before you ask. ChatGPT monitors your calendar, reads your emails and offers recommendations without being asked.
Soon comes “autopilot AI,” where AI makes purchases for you with minimal input from you. “Order flowers for my anniversary next week.” ChatGPT checks your calendar, remembers preferences, processes payment and confirms delivery.
Each phase adds convenience but gives you less control.
The manipulation problem
AI’s responses create what researchers call an “advice illusion.” When ChatGPT suggests three hotels, you don’t see them as ads. They feel like recommendations from a knowledgeable friend. But you don’t know whether those hotels paid for placement or whether better options exist that ChatGPT didn’t show you.
Traditional advertising is something most people have learned to recognize and dismiss. But AI recommendations feel objective even when they’re not. With one-tap purchasing, the entire process happens so smoothly that you might not pause to compare options.
OpenAI isn’t alone in this race. In the same month, Google announced its competing protocol, AP2. Microsoft, Amazon and Meta are building similar systems. Whoever wins will be in position to control how billions of people buy things, potentially capturing a percentage of trillions of dollars in annual transactions.
What we’re giving up
This convenience comes with costs most people haven’t thought about.
Privacy: For AI to suggest restaurants, it needs to read your calendar and emails. For it to buy flowers, it needs your purchase history. People will be trading total surveillance for convenience.
Choice: Right now, you see multiple options when you search. With AI as the middleman, you might see only three options ChatGPT chooses. Entire businesses could become invisible if AI chooses to ignore them.
Power of comparing: When ChatGPT suggests products with one-tap checkout, the friction that made you pause and compare disappears.
It’s happening faster than you think
ChatGPT reached 800 million weekly users by September 2025, growing four times faster than social media platforms did. Major retailers began using OpenAI’s Agentic Commerce Protocol within days of its launch.
History shows people consistently underestimate how quickly they adapt to convenient technologies. Not long ago most people wouldn’t think of getting in a stranger’s car. Uber now has 150 million users.
Convenience always wins. The question isn’t whether AI shopping will become mainstream. It’s whether people will keep any real control over what they buy and why.
What you can do
The open internet gave people a world of information and choice at their fingertips. The AI revolution could take that away. Not by forcing people, but by making it so easy to let the algorithm decide that they forget what it’s like to truly choose for themselves. Buying things is becoming as thoughtless as sending a text.
In addition, a single company could become the gatekeeper for all digital shopping, with the potential for monopolization beyond even Amazon’s current dominance in e-commerce. We believe that it’s important to at least have a vigorous public conversation about whether this is the future people actually want.
Here are some steps you can take to resist the lure of convenience:
Question AI suggestions. When ChatGPT suggests products, recognize you’re seeing hand-picked choices, not all your options. Before one-tap purchases, pause and ask: Would I buy this if I had to visit five websites and compare prices?
Review your privacy settings carefully. Understand what you’re trading for convenience.
Talk about this with friends and family. The shift to AI shopping is happening without public awareness. The time to have conversations about acceptable limits is now, before one-tap purchasing becomes so normal that questioning it seems strange.
The invisible price tag
AI will learn what you want, maybe even before you want it. Every time you tap “Buy now,” you’re training it—teaching it your patterns, your weaknesses, what time of day you impulse buy.
Our warning isn’t about rejecting technology. It’s about recognizing the trade-offs. Every convenience has a cost. Every tap is data. The companies building these systems are betting you won’t notice, and in most cases they’re probably right.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
Citation:
OpenAI has slipped shopping into ChatGPT users’ chats—here’s why that matters (2025, October 20)
retrieved 20 October 2025
from https://techxplore.com/news/2025-10-openai-chatgpt-users-chats.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
Tech
Need One Pair for Hiking, Traveling, and Working Out? Try Gravel Running Shoes
HOKA’s max-stacked Rocket X Trail combines road race shoe energy with boosted grip from a 3-mm lugged outsole. If you’re looking for a fast shoe to go on the attack, this is it. It’s also fantastic for all round comfort. In testing, I laced up the Rocket X Trail and ran 3 hours (just short of 19 miles) fresh out of the box, across roads, forest gravel trails, some grass and through some serious water. It delivered efficiency and energy whether I was moving at marathon pace or with heavier, tired, ragged footfalls in the latter miles.
The rockered, supercritical midsole uses HOKA’s liveliest foam, similar to those you find in its race-ready road shoes, along with a carbon plate. That combines for a really fun ride that’s smooth, springy and fast and really consistent. It’s also highly cushioned, so you will sacrifice a lot of ground feel for that big stack springy softness. It’s also less stable over very lumpy terrain. But on open, flat, runnable mixed terrain, it’s excellent.
The lightweight uppers have a race-shoe-ready feel and after running through ankle-deep flooded sections, they shed water really quickly. This is a pricey road-to-trail shoe, it’s versatile and there’s plenty of winter road potential, too.
| Specs | |
|---|---|
| Weight | 9.45 oz |
| Heel-to-toe drop | 6 mm |
| Lug depth | 3 mm |
Tech
If a Garmin Is Too Expensive, Consider Suunto’s Latest Adventure Watch
It’s always pleasing to see an array of physical buttons, and you get sizable ones too. You’re not going to miss these wide flat ones even when picking the pace up. The silicone strap has a nice stretch to it and while the button clasp is a bit awkward to get into place, this watch does not budge.
Suunto has jumped on the flashlight trend, with an LED light strip sat on the front of the case. You can adjust brightness levels and there’s SOS and alert modes to emit a very noticeable pulsating light pattern. This is a light I found useful rooting around indoors as well as on nighttime outings.
The biggest change is the introduction of a 1.5-inch, 466 x 466 AMOLED display. This replaces the dull, albeit very visible, memory-in-pixel (MIP) display. Suunto also ditched the solar charging that did require spending a significant amount of time outside to reap its battery benefits.
Adding AMOLED screens to outdoor watches has been contentious. The older MIP displays are just more power-efficient. The Vertical 2 is down by about 10 days from the older Vertical for what Suunto calls daily use.
Still, even if you’re putting its tracking and mapping features to use, you’re not going to be reaching for the charger every few days. After two hours of tracking in optimal GPS mode, the battery only dropped by 2 to 3 percent. The battery drop outside of tracking is also small and the standby performance is excellent as well.
Software Updates
Photograph: Michael Sawh
A more streamlined set of smartwatch features helps reserve battery for when it really matters. Unfortunately, I probably got better battery life because you don’t get phone notifications or responses if it’s paired to an iPhone instead of an Android. There’s also no onboard music player, but you do get a pretty slick set of music playback controls that are accessible during tracking.
Tech
Electronic health records are still creating issues for patients | Computer Weekly
Every NHS trust in England needs an electronic patient record (EPR) system in place by March 2026, as part of a government push to digitise the healthcare system.
In many ways, this is long overdue: some trusts have still been using pen-and-paper record-keeping until very recently.
EPRs have the potential to massively improve efficiency in the NHS. If working properly, they allow doctors to keep all of their records in one place, speed up prescribing and diagnostics, and make it easier for patients to access their own health information.
But these roll-outs have not been without problems. Concerns have been raised about how far these benefits can actually be realised. Some NHS trusts have experienced issues with integrating new systems and training staff on how to use them.
In the extreme, there have been reports of EPRs creating new problems for hospitals, with evidence suggesting these systems may have contributed to serious harm and even deaths among patients.
NHS trusts have been put in charge of procuring their own EPRs, meaning there are numerous different technology companies involved. Some providers of these systems are large US firms. This includes Oracle Health, provided by the Larry Ellison-led tech giant, and Epic, a tech firm based in Wisconsin.
Contracts can run into nine figures: Guy’s and St Thomas’, a trust in South London, launched a £450m system from Epic in late 2023. Some parts of the NHS have been using them for more than a decade, but a handful are still set to miss the government’s March deadline.
Data access
Pritesh Mistry is a fellow at the King’s Fund, where he researches the impact of digital transformation in the NHS. He says it has had “both positive and negative impacts”.
“In the last few years, we’ve seen doubling down on the focus around digital records,” says Mistry. These are now in place in more than 90% of all trusts, and every GP practice.
“That means we’ve now got [new] data that’s within the healthcare system, which allows us to do other things, like treat populations, and understand and track patient safety,” he says.
Despite this, he cautions some patients are still struggling to get hold of their own data.
“We’ve got a lot of data that’s in silos,” says Mistry. “It doesn’t flow. That’s the biggest challenge: making the data accessible and usable for patients and healthcare professionals to be able to provide care in a way that is joined up and meets with modern expectations.”
He says complaints with new technology haven’t just come from patients.
“We need to recognise that staff are really frustrated,” says Mistry. “Software often crashes. Computers are really slow, and technology adds to their workload, instead of simplifying things.” He caveats that some parts of the NHS are better than others on this.
Safeguarding patient data
Mistry adds that there are safeguards in place to ensure patient data isn’t ending up where it shouldn’t be – such as through data protection rules and procurement requirements.
However, he warns that “we need to make sure we move with the times in terms of what technology is available”. Mistry is more concerned about medical staff inadvertently putting personal information into a large language model, for instance.
“Digital exclusion remains a barrier as well,” he says, adding that these systems have the potential to widen inequalities in healthcare. Those less able to use new technology might struggle to access their records.
“People tend to assume it’s old people [who are most impacted], but that isn’t necessarily true,” says Mistry, instead highlighting the impact of poverty and deprivation, with some still unable to afford internet access.
He argues the NHS should be working to meet people where they are, and provide more “tailored” technology services.
Patient safety
Nick Woodier is a doctor and investigator at the Health Services Safety Investigations Body (HSSIB), which looks into issues with healthcare in the UK. He sees problems arising from how EPRs are deployed by trusts, especially when medical staff overestimate their capabilities.
He uses the example of prescribing medicines: “There’s an assumption that these electronic prescribing systems will stop you [from] doing something catastrophic.”
But this isn’t always the case. In one investigation, the HSSIB found a child had been prescribed nearly 10 times the recommended dose of an anti-coagulant medication, with doctors having assumed the EPR would flag an issue. The child ended up with a bleed on their brain.
Woodier also worries hospitals are not always picking up on when these systems are at fault.
“We will often see where incidents have happened and the contribution of the electronic system has not been recognised,” he says.
Woodier sees this as coming from a culture which prefers to put the blame for safety failures on individuals.
A 2024 investigation by the BBC found there were more than 126 instances of serious harm registered by NHS trusts across 31 trusts, including three deaths related to EPR problems.
The HSSIB has also encountered problems from patients being unable to access their digital records.
“We’ve seen in general practice, for example, some patients telling us that they’ve gone without care – because in their mind, they thought the only way they could access their GP was to fill in an electronic form,” says Woodier.
A spokesperson for NHS England says EPRs are “already having a significant impact on improving safety and care for patients”, for instance, by helping to identify conditions such as sepsis, and preventing medication errors.
“They have replaced outdated and often less-safe paper-based systems, and we are working closely with NHS trusts to ensure they are implemented safely alongside other systems with appropriate training – and are used to the highest quality and safety standards,” the spokesperson adds.
Interoperability
The EPR roll-out has also been criticised for problems with “interoperability” – the ability of different programs and modes of data collection to converse with each other. The patchwork of different systems used by different trusts means data stored in one system might not be useful for a system used by a different part of the NHS.
Woodier says this often happens in communications between hospitals and GP surgeries. This can involve someone manually inputting information from one system to another, which can create risks when data is not being transferred properly, or is missed completely.
“When you introduce a manual operation, that risk increases,” he warns. “The odds are that at some point, somebody won’t do the right thing, because that’s the reality of being human.”
Alex Lawrence, a fellow at the Health Foundation, describes interoperability as a “significant challenge”, which the NHS and technology companies have been “grappling with for a really long time”.
“Some trusts have found it much harder to access their own EPR data than they anticipated, because of where that data is stored,” she adds, referring to research the organisation carried out in 2024.
“If it’s taking you days to pull the data that you need, then it’s already not going to be useful for a lot of the purposes that you might want it for.”
However, Lawrence adds that there have been some steps made in the right direction, notably with the Data (Use and Access) Act, which was passed last year.
“The government is making information standards mandatory for EPR providers, as well as trusts, with the Secretary of State potentially having more powers to enforce those standards,” she says.
The longer term
Going forward, Lawrence would like to see a system involving “patients being empowered with access to their own data, and as far as appropriate, clinicians being able to see all of the history that they need for their patients”.
In an ideal system, different parts of the healthcare system would be able to “share a patient’s data where necessary and appropriate, in an easy and timely way”.
She says they have the “potential to offer enormous value”, but much of their functionality is going unused. “What our qualitative research suggested was that a lot of these systems are still functioning as digital notebooks,” says Lawrence.
Matthew Taylor is the head of the NHS Confederation and NHS Providers, membership bodies for healthcare organisations.
“NHS leaders say the gap between trusts on digital maturity is still stark – and it’s shaping how quickly organisations can move to modern EPRs,” he says.
This gap – combined with the organisational complexity of the healthcare system – means interoperability has “long been a thorn in the NHS’s side”.
Taylor adds that EPRs are not a “once-and-done” job, and argues they will result in savings in the long term, but that it may take around five years to see the benefits.
“Hospitals are housing a huge amount of paper records, and the cost of storing, retrieving and managing those records can run into millions of pounds each year,” he says.
These systems are part of a larger picture, and one facet of the conversation, around the use of artificial intelligence in the NHS. AI models for areas such as research and diagnostics will require extensive and standardised medical data.
Mistry warns these AI tools operate on the basis of “garbage in, garbage out”.
“There is a risk that we roll out AI tools without the underpinning data quality it needs,” he says, adding that this could exacerbate inequalities or biases from using AI.
As Woodier puts it: “We’ve got organisations who are still using archaic computers, have got infrastructure that’s not working, are still on old web systems, or have EPRs that don’t talk to each other. A few [trusts] don’t have EPRs.
“So, actually, are we trying to run before we’ve even managed to walk?”
-
Politics1 week agoWhat are Iran’s ballistic missile capabilities?
-
Business6 days agoIndia Us Trade Deal: Fresh look at India-US trade deal? May be ‘rebalanced’ if circumstances change, says Piyush Goyal – The Times of India
-
Politics1 week agoUS arrests ex-Air Force pilot for ‘training’ Chinese military
-
Business7 days agoAttock Cement’s acquisition approved | The Express Tribune
-
Business1 week agoHouseholds set for lower energy bills amid price cap shake-up
-
Fashion7 days agoPolicy easing drives Argentina’s garment import surge in 2025
-
Sports6 days agoLPGA legend shares her feelings about US women’s Olympic wins: ‘Gets me really emotional’
-
Fashion7 days agoTexwin Spinning showcasing premium cotton yarn range at VIATT 2026
