Exclusive | United Airlines Is Banning AI Bots From Booking Flights – WSJ

United Airlines said in November 20206 that, just like Amazon, it will prevent personal artificial-intelligence assistants from accessing its systems to book tickets. Now, while merchants such as these (and major banks) present security as the key reason fot blocking agents, some industry observers think that their action are more to with economics. Tomas Tunguz, a venture capital guy, sets out the case for this view by noting that in modern e-commerce, the base product breaks even whereas the “add ons” provide the profits. He uses three examples to make this point:

Amazon’s sponsored lists generate  $34 billion at a 50% margin, which is 100% of their retail operating income;

United’s $5 billion in upgrades plus $3 billion of credit card, loyalty and insurance fees are more than is operating income;

Dell’s 14% gross margin does not cover $15 billion opex, where as ProSupport and attached services more than 120% of operating income.

An AI agent is a buyer skips the sponsored listings, declines Economy Plus unless told to unchecks the extended warranty on a Dell Latitude. Merchants built on loss leaders now have to either raise the base prices to make a living, start charging bots to make up for the difference or come up with some other incentives to persuade bots to take out the airlines’ insurance, the banks’ credit cards and the retailers’ impulse purchases.

, charge an explicit agent checkout fee, or invent wholesale incentives that persuade bots to upsell.
Until the funnel is re-priced, the sign on the checkout window stays the same : no bots allowed.

Note that these business models are very different to Shopify’s model, which is collecting three per cent of every dollar they process, irrespective of whether it is a person or an agent pressing the buttons.

AI Companies Really, Really Want to Buy Your Work Files and Emails | WIRED

xxx

Big AI has two bottomless needs: compute and data. While the industry is very publicly making giant investments in the former, data gathering is just as important.
MOST POPULAR

DEALS
Can’t Wait for Black Friday? Check Out These Live Deals
BY SIMON HILL

POLITICS
‘A Servant Class, Not a Working Class’: Stephen Miller’s Biographer on What Comes Next
BY VITTORIA ELLIOTT

BUSINESS
‘Bring Back THE Boss’: Nvidia Employees Petition for Jensen Huang’s Return
BY PARESH DAVE

SECURITY NEWS
The Microsoft Outage Is a Warning for the World
BY LILY HAY NEWMAN
Foody is ready when I bring up privacy concerns. When a business streams the data to Mercor, the company strips out any personal identifying information, or PII. This de-identification can be verified, he says. Like all the people in data acquisition I spoke to, he felt that technology has solved the privacy issues of buying all that information.

Not surprisingly, Micro1 CEO Ali Ansari agrees that the data-buying market will soon approach that of AI’s compute infrastructure. He says that today, labs, enterprises, and robotics companies pay about $40 to $50 billion a year to get data.

From: AI Companies Really, Really Want to Buy Your Work Files and Emails | WIRED.

xxx

How Zero-Knowledge Proofs Could Fortify EU’s Digital ID – IEEE Spectrum

xxx

Millions of European Union citizens should soon have access to digital wallets, as member states roll out their implementations of the European Digital Identity Regulation ahead of an end-of-the-year deadline. But the digital-identity wallets, which are intended to provide a secure and private place to store identifying information such as bank-account information, diplomas, and government IDs, such as driver’s licenses, could leave behind a trail of cryptographic clues that makes it possible to recognize someone across services. Cryptographers warn that the technology protecting individual privacy may not yet go far enough to stop people from being tracked.

Advances in cryptographic techniques since the law was passed in 2024 means this problem could be solved. However, the law doesn’t require that these techniques be used, and cryptographers that IEEE Spectrum spoke with are split as to whether the wallets should be rolled out at all without such a requirement in place, and whether they’ll be secure enough without it.

With digital wallets, when someone is asked to provide their ID, a digital, verifiable form of the ID can be transmitted, via a QR code for example. When scanned with a verification app, it would show basic identifying user information, according to Anna Lysyanskaya, a professor of computer science at Brown University.

The EU’s digital-identity regulation already includes a requirement for selective disclosure, which allows people to share only the attributes requested by the person checking identification rather than handing over an entire document. A bar checking someone’s age, for example, could request a simple confirmation that a person is over 18 and not need their full name or address..

Tracking Risks in EU Digital Wallets

When the person who requested the ID scans it, they still record some personal information, such as the digital signature and identifiers that are required to check the ID. This additional information can serve as a way of tracking a user even if the attributes shared are quite common. Two requesters could share this information to show they had interacted with the same person, says Johannes Schönrich, the director at the FIM Research Center for Information Management at the Augsburg Technical University of Applied Sciences, in Germany.

If two parties can share that information, that would be a problem: EU legislation specifies that the wallets should also be unlinkable, which means that “tracking a user by means of [these] digital attestations should not be possible,” says Schönrich.

One way to address the problem is to issue multiple versions of a user’s credentials, each with different cryptographic information attached, says Lysyanskaya. The wallet could keep track of which certificates have already been used, ensuring that a different one is presented during each interaction.

“If you allow a service provider to say, Look, I’m not supporting the Google protocol, I want a traditional verification, you essentially allow the service provider to do a kind of downgrade.”
—JAAP-HENK HOEPMAN, RADBOUD UNIVERSITY

This approach has limitations. If the issuer—the government department that creates and digitally verifies that the credential is authentic—can identify who a credential was originally issued to, it could potentially help link the different versions back to the same person, according to Lysyanskaya.

“Now anybody you’ve ever interacted with has some kind of way of tracking you,” she says.

“If this kind of interaction becomes really ubiquitous…then unless we place a tremendous amount of trust in both issuers and verifiers to guard our privacy, we may encounter a situation where everybody knows at all times what everybody else is up to, by just leaving these little digital crumbs of information about yourself.”

Lysyanskaya was one of 16 cryptographers who were invited to provide feedback for the European Commission when the law was passed in 2024. They recommended using anonymous credentials, which use zero-knowledge proofs to let people demonstrate their identity or eligibility without revealing unnecessary personal information.

How Zero-Knowledge Proofs Work

Zero-knowledge proofs allow one party to convince another that something is true without needing to send any additional information beyond the fact that it is true. For example, they would allow someone to demonstrate that they hold a valid government-issued credential, or meet a particular requirement, without revealing the underlying personal information or creating the same linkable record of the interaction, says Lysyanskaya.

If these two people, the person checking the ID and the person being asked for it ever meet again, there will be no record of any earlier encounters.

The service provider could still verify that the credential is authentic, has not expired or been revoked, and has not been manipulated. However, it would receive less information than it does under traditional verification methods.

“You learn exactly what you need, but not a single bit more about anything that might accidentally be disclosed but is not that necessary,” says Schönrich.

Zero-knowledge proofs have existed in cryptography since the 1980s, but teams of researchers have only recently developed techniques to run them on more-constrained devices like smartphones. Shortly after the EU’s digital-wallet regulations were passed in 2024, for example, a research team at Google found a way to generate zero-knowledge proofs using credentials signed with the elliptic-curve digital-signature algorithm, which generates and authenticates digital signatures efficiently enough to be used in a digital wallet.

From: How Zero-Knowledge Proofs Could Fortify EU’s Digital ID – IEEE Spectrum.

xxx

Post | LinkedIn

Tomas Tunguz, a venture capital guy, said this clearly in a post on LinkedIn. While United and Amazon talk about security, perhaps there is an underlying economic reason.

United Airlines banned AI agents from booking flights this week, reserving the right to cancel any ticket a bot buys. Amazon blocked third-party shopping agents from its store.

Both cite security. The real reason sits on their income statements : in modern e-commerce, the base product breaks even.

Ads, upgrades & warranties supply the profit.

– Amazon : $588b of retail revenue, but its $68.6b of sponsored listings generate $34.3b at a 50% margin. 100% of Amazon’s $34.4b in retail operating income.
– United : $4.8b of seat & bag upgrades plus $3.2b of credit card, loyalty & trip insurance fees top its entire $4.7b of operating income.
– Dell : hardware at a 13.7% gross margin does not cover $14.6b of opex. ProSupport & attached services at a 44.8% margin generate $10.4b of gross profit, 127% of Dell’s $8.1b in operating income.

Shopify runs the opposite model, collecting a 3.1% toll on every dollar processed. Whether a human or an agent clicks buy, Shopify takes the same cut. That is why it opened checkout to browser agents across every eligible store.

An AI agent is a buyer on a mission. It skips the sponsored listings, declines Economy Plus unless told & unchecks the extended warranty on a Dell Latitude.

Merchants built on loss leaders now face three choices : raise base prices, charge an explicit agent checkout fee, or invent wholesale incentives that persuade bots to upsell.
Until the funnel is re-priced, the sign on the checkout window stays the same : no bots allowed.

From: Post | LinkedIn.

xxx

Kalshi cuts ties to Donald Trump Jr after Democrats’ midterm sweep

xxx

Kalshi has said it is severing ties with Donald Trump Jr a week after Democrats swept to victory in the midterm elections, pledging to investigate ties between the US president’s family and corporate America.

From: Kalshi cuts ties to Donald Trump Jr after Democrats’ midterm sweep.

xxx

Personal AI assistants could soon reshape the digital economy

xxx

Prolonged exposure appears to build that trust. About 40% of Instinct users link a credit card within three weeks. Instinct and Meta also work with Stripe, a payments company, to offer a digital wallet that allows users to make purchases without handing over their banking details directly, which may reassure some.

From: Personal AI assistants could soon reshape the digital economy.

xxx

Australia opens myID to business, brings private providers into govt system | Biometric Update

xxx

The opening completes the phased rollout established under the Digital ID Act 2024 and introduces a new commercial dynamic. The government-backed myID gives businesses access to an established national identity ecosystem, but that reach will come at a price, while state-issued mobile driver’s licenses (mDLs) and other verifiable credentials offer an alternative that critics say will be cheaper to accept. Private identity providers, meanwhile, gain access to government customers, although restrictions will limit which services they can support.

From: Australia opens myID to business, brings private providers into govt system | Biometric Update.

xxx

Pentagon seeks new identity system as modernization gaps persist | Biometric Update

xxx

The Federation Hub allows different Defense agencies, military services, and approved partners to accept authentication from trusted identity providers without requiring users to establish separate credentials for every system.
Confirming someone’s identity, though, does not automatically authorize access to sensitive information. Those decisions may also depend on factors such as organizational affiliation, security clearance, training and assigned duties.
DISA says the current environment relies heavily on coarse-grained role-based access control rather than the dynamic, attribute-based authorization needed for its zero-trust architecture.

From: Pentagon seeks new identity system as modernization gaps persist | Biometric Update.

xxx

Create your own website, free on WordPress.com.Create your own website.
Start free