Smart contracts beyond the age of innocence

Why have Bitcoin, with its distributed consistent ledger, and now Ethereum with its support for fully fledged “smart contracts,” captured the imagination of so many people, both within and beyond the tech industry? The promise to replace obscure stores of information and arcane contract rules – with their inefficient, ambiguous, and primitive human interpretations – with publicly visible decentralized ledgers reflects the growing technological zeitgeist in their guarantee that all participants would know and be able to foresee the consequences of both their own actions and the actions of all others. The precise specification of contracts as code, with clauses automatically executed depending on certain sets of events and permissible user actions, represents for some a true state of utopia.

Regardless of one’s views on the potential for distributed ledgers, one of the most notable innovations that smart contracts have enabled thus far is the idea of a DAO (Decentralized Autonomous Organization), which is a specific type of investment contract, by which members individually contribute value that then gets collectively invested under some governance model.  In truly transparent fashion, the details of this governance model, including who can vote and how many votes are required for a successful proposal, are all encoded in a smart contract that is published (and thus globally visible) on the distributed ledger.

Today, this vision met a serious stumbling block: a “bug” in the contract of the first majorly successful DAO (which broke records by raising 11 million ether, the equivalent of 150 million USD, in its first two weeks of operation) allowed third parties to start draining its funds, and to eventually make off with 4% of all ether. The immediate response of the Ethereum and DAO community was to suspend activity – seemingly an anathema for a ledger designed to provide high resiliency and availability – and propose two potential solutions: a “soft-fork” that would impose additional rules on miners in order to exclude all future transactions that try to use the stolen ether, or, more drastically (and running directly contrary to the immutability of the ledger),  a “hard-fork” that would roll back the transactions in which the attack took place, in addition to the many legitimate transactions that took place concurrently.  Interestingly, a variant of the bug that enabled the hack was known to and dismissed by the creators of the DAO (and the wider Ethereum community).

While some may be surprised by this series of events, Maurice Wilkes, designer of the EDSAC, one of the first computers, reflected that “[…] the realization came over me with full force that a good part of the remainder of my life was going to be spent in finding errors in my own programs.” It is not the case that because a program is precisely defined it is easy to foresee what it will do once executed on its own under the control of users.  In fact, Rice’s theorem explicitly states that it is not possible in general to show that the result of programs, and thus smart contracts, will have any specific non-trivial property.

This forms the basis on which modern verification techniques operate: they try to define subsets of programs for which it is possible to prove some properties (e.g., through typing), or attempt to prove properties in a post-hoc way (e.g., through verification), but under the understanding that they may fail in general.  There is thus no scientific basis on which one can assert generally that smart contracts can easily provide clarity into and foresight of their consequences.

The unfolding story of the DAO and its consequences for the Ethereum community offers two interesting insights. First, as a sign that the field is maturing, there is an explicit call for understanding the computational space of safe contracts, and contracts with foreseeable consequences. Second, it suggests the need for smart contracts protecting significant assets to include external, possibly social, mechanisms in order to unlock significant value transfers. The willingness of exchanges to suspend trading and of the Ethereum developers to suggest a hard-fork is a last-resort example of such a social mechanism. Thus, politics – the discipline of collective management – reasserts itself as having primacy over human affairs.

Exceptional access provisions in the Investigatory Powers Bill

The Investigatory Powers Bill, being debated in Parliament this week, proposes the first wide-scale update in 15 years to the surveillance powers of the UK law-enforcement and intelligence agencies.

The Bill has several goals: to consolidate some existing surveillance powers currently either scattered throughout other legislation or not even publicly disclosed, to create a wide range of new surveillance powers, and to change the process of authorisation and oversight surrounding the use of surveillance powers. The Bill is complex and, at 245 pages long, makes scrutiny challenging.

The Bill has had its first and second readings in the House of Commons, and has been examined by relevant committees in the Commons. The Bill will now be debated in the ‘report stage’, where MPs will have the chance to propose amendments following committee scrutiny. After this it will progress to a third reading, and then to the House of Lords for further debate, followed by final agreement by both Houses.

So far, four committee reports have been published examining the draft Bill, from the Intelligence and Security Committee of Parliament, the joint House of Lords/House of Commons committee specifically set up to examine the draft Bill, the House of Commons Science and Technology committee (to which I served as technical advisor) and the Joint Committee on Human Rights.

These committees were faced with a difficult task of meeting an accelerated timetable for the Bill, with the government aiming to have it become law by the end of 2016. The reason for the haste is that the Bill would re-instate and extend the ability of the government to compel companies to collect data about their users, even without there being any suspicion of wrongdoing, known as “data retention”. This power was previously set out in the EU Data Retention Directive, but in 2014 the European Court of Justice found it be unlawful.

Emergency legislation passed to temporarily permit the government to continue their activities will expire in December 2016 (but may be repealed earlier if an appeal to the European Court of Justice succeeds).

The four committees which examined the Bill together made 130 recommendations but since the draft was published, the government only slightly changed the Bill, and only a few minor amendments were accepted by the Public Bills committee.

Many questions remain about whether the powers granted by the Bill are justifiable and subject to adequate oversight, but where insights from computer security research are particularly relevant is on the powers to grant law enforcement the ability to bypass normal security mechanisms, sometimes termed “exceptional access”.

Continue reading Exceptional access provisions in the Investigatory Powers Bill

International Comparison of Bank Fraud Reimbursement: Customer Perceptions and Contractual Terms

Terms and Conditions (T&C) are long, convoluted, and are very rarely actually read by customers. Yet when customers are subject to fraud, the content of the T&Cs, along with national regulations, matter. The ability to revoke fraudulent payments and reimburse victims of fraud is one of the main selling points of traditional payment systems, but to be reimbursed a fraud victim may need to demonstrate that they have followed security practices set out in their contract with the bank.

Security advice in banking terms and conditions vary greatly across the world. Our study’s scope included Europe (Cyprus, Denmark, Germany, Greece, Italy, Malta, and the United Kingdom), the United States, Africa (Algeria, Kenya, Nigeria, and South Africa), the Middle East (Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Oman, Palestine, Qatar, Saudi Arabia, UAE and Yemen), and East Asia (Singapore). Out of 30 banks’ terms and conditions studied, 26 give more or less specific advice on how you may store your PIN. The advice varies from “Never writing the Customer’s password or security details down in a way that someone else could easily understand” (Arab Banking Corp, Algeria), “If the Customer makes a written record of any PIN Code or security procedure, the Customer must make reasonable effort to disguise it and must not keep it with the card for which it is to be used” (National Bank of Kenya) to “any record of the PIN is kept separate from the card and in a safe place” (Nedbank, South Africa).

Half of the T&Cs studied give advice on choosing and changing one’s PIN. Some banks ask customers to immediately choose a new PIN when receiving a PIN from the bank, others don’t include any provision for customers to change their PIN. Some banks give specific advice on how to choose a PIN:

When selecting a substitute ATM-PIN, the Customer shall refrain from selecting any series of consecutive or same or similar numbers or any series of numbers which may easily be ascertainable or identifiable with the Customer…

— OCBC, Singapore

Only 5 banks give specific advice about whether you are allowed to re-use your PIN on other payment cards or elsewhere. There is also disagreement about what to do with the PIN advice slip, with 7 banks asking the customer to destroy it.

Some banks also include advice on Internet security. In the UK, HSBC for example demands that customers

always access Internet banking by typing the address into the web browser and use antivirus, antispyware and a personal firewall. If accessing Internet banking from a computer connected to a LAN or a public Internet access device or access point, they must first ensure that nobody else can observe, copy or access their account. They cannot use any software, such as browsers or password managers, to record passwords or other security details, apart from a service provided by the bank. Finally, all security measures recommended by the manufacturer of the device being used to access Internet banking must be followed, such as using a PIN to access a mobile device.

— HSBC, UK

Over half of banks tell customers to use firewalls and anti-virus software. Some even recommend specific commercial software, or tell customers how to find some:

It is also possible to obtain free anti-virus protection. A search for `free anti-virus’ on Google will provide a list of the most popular.

— Commercial International Bank, Egypt

In the second part of our paper, we investigate the customers’ perception of banking T&Cs in three countries: Germany, the United States and the United Kingdom. We present the participants with 2 real-life scenarios where individuals are subject to fraud, and ask them to decide on the outcome. We then present the participants with sections of T&Cs representative for their country and ask them then to re-evaluate the outcome of the two scenarios.

Question DE UK US
Scenario 1: Card Loss 41.5% 81.5% 76.8%
Scenario 1: Card Loss after T&Cs 70.7% 66.7% 96.4%
Scenario 2: Phishing 31.7% 37.0% 35.7%
Scenario 2: Phishing after T&Cs 43.9% 46.3% 42.9%

The table above lists the percentage of participants that say that the money should be returned for each of the scenarios. We find that in all but one case, the participants are more likely to have the protagonist reimbursed after reading the terms and conditions. This is noteworthy – our participants are generally reassured by what they read in the T&Cs.

Further, we assess the participants’ comprehension of the T&Cs. Only 35% of participants fully understand the sections, but the regional variations are large: 45% of participants in the US fully understanding the T&Cs but only 22% do so in Germany. This may indeed be related to the differences in consumer protection laws between the countries: In the US, Federal regulations give consumers much stronger protections. In Germany and the UK (and indeed, throughout Europe under the EU’s Payment Service Directive), whether a victim of fraud is reimbursed depends on if he/she has been grossly negligent – a term that is not clearly defined and confused our participants throughout.

 

International Comparison of Bank Fraud Reimbursement: Customer Perceptions and Contractual Terms by Ingolf Becker, Alice Hutchings, Ruba Abu-Salma, Ross Anderson, Nicholas Bohm, Steven J. Murdoch, M. Angela Sasse and Gianluca Stringhini will be presented at the Workshop on the Economics of Information Security (WEIS), Berkeley, CA USA, 13–14 June 2016.

Adblocking and Counter-Blocking: A Slice of the Arms Race

anti-adblocking message from WIRED
If you use an adblocker, you are probably familiar with messages of the kind shown above, asking you to either disable your adblocker, or to consider supporting the host website via a donation or subscription. This is the battle du jour in the ongoing adblocking arms race — and it’s one we explore in our new report Adblocking and Counter-Blocking: A Slice of the Arms Race.

The reasons for the rising popularity of adblockers include improved browsing experience, better privacy, and protection against malvertising. As a result, online advertising revenue is gravely threatened by adblockers, prompting publishers to actively detect adblock users, and subsequently block them or otherwise coerce the user to disable the adblocker — practices we refer to as anti-adblocking. While there has been a degree of sound and fury on the topic, until now we haven’t been able to understand the scale, mechanism and dynamics of anti-adblocking. This is the gap we have started to address, together with researchers from the University of Cambridge, Stony Brook University, University College London, University of California Berkeley, Queen Mary University of London and International Computer Science Institute (Berkeley). We address some of these questions by leveraging a novel approach for identifying third-party services shared across multiple websites to present a first characterization of anti-adblocking across the Alexa Top-5K websites.

We find that at least 6.7% of Alexa Top-5K websites employ anti-adblocking, with the practices finding adoption across a diverse mix of publishers; particularly publishers of “General News”, “Blogs/Wiki”, and “Entertainment” categories. It turns out that these websites owe their anti-adblocking capabilities to 14 unique scripts pulled from 12 unique domains. Unsurprisingly, the most popular domains are those that have skin in the game — Google, Taboola, Outbrain, Ensighten and Pagefair — the latter being a company that specialises in anti-adblocking services. Then there are in-house anti-adblocking solutions that are distributed by a domain to client websites belonging to the same organisation: TripAdvisor distributes an anti-adblocking script to its eight websites with different country code top-level domains, while adult websites (all hosted by MindGeek) turn to DoublePimp. Finally, we visited a sample website for each anti-adblocking script via AdBlock Plus, Ghostery and Privacy Badger, and discovered that half of the 12 anti-adblocking suppliers are counter-blocked by at least one adblocker — suggesting that the arms race has already entered the next level.

It is hard to say how many levels deeper the adblocking arms race might go. While anti-adblocking may provide temporary relief to publishers, it is essentially band-aid solution to mask a deeper issue — the disequilibrium between ads (and, particularly, their behavioural tracking back-end) and information. Any long term solution must address the reasons that brought users to adblockers in the first place. In the meantime, as the arms race continues to escalate, we hope that studies such as ours will bring transparency to this opaque subject, and inform policy that moves us out of the current deadlock.

 

“Ad-Blocking and Counter Blocking: A Slice of the Arms Races” by Rishab Nithyanand, Sheharbano Khattak, Mobin Javed, Narseo Vallina-Rodriguez, Marjan Falahrastegar, Julia E. Powles, Emiliano De Cristofaro, Hamed Haddadi, and Steven J. Murdoch. arXiv:1605.05077v1 [cs.CR], May 2016.

This post also appears on the University of Cambridge Computer Laboratory Security Group blog, Light Blue Touchpaper.

On the hunt for Facebook’s army of fake likes

As social networks are increasingly relied upon to engage with people worldwide, it is crucial to understand and counter fraudulent activities. One of these is “like farming” – the process of artificially inflating the number of Facebook page likes. To counter them, researchers worldwide have designed detection algorithms to distinguish between genuine likes and artificial ones generated by farm-controlled accounts. However, it turns out that more sophisticated farms can often evade detection tools, including those deployed by Facebook.

What is Like Farming?

Facebook pages allow their owners to publicize products and events and in general to get in touch with customers and fans. They can also promote them via targeted ads – in fact, more than 40 million small businesses reportedly have active pages, and almost 2 million of them use Facebook’s advertising platform.

At the same time, as the number of likes attracted by a Facebook page is considered a measure of its popularity, an ecosystem of so-called “like farms” has emerged that inflate the number of page likes. Farms typically do so either to later sell these pages to scammers at an increased resale/marketing value or as a paid service to page owners. Costs for like farms’ services are quite volatile, but they typically range between $10 and $100 per 100 likes, also depending on whether one wants to target specific regions — e.g., likes from US users are usually more expensive.

Screenshot from http://www.getmesomelikes.co.uk/
Screenshot from http://www.getmesomelikes.co.uk/

How do farms operate?

There are a number of possible way farms can operate, and ultimately this dramatically influences not only their cost but also how hard it is to detect them. One obvious way is to instruct fake accounts, however, opening a fake account is somewhat cumbersome, since Facebook now requires users to solve a CAPTCHA and/or enter a code received via SMS. Another strategy is to rely on compromised accounts, i.e., by controlling real accounts whose credentials have been illegally obtained from password leaks or through malware. For instance, fraudsters could obtain Facebook passwords through a malicious browser extension on the victim’s computer, by hijacking a Facebook app, via social engineering attacks, or finding credentials leaked from other websites (and dumped on underground forums) that are also valid on Facebook.

Continue reading On the hunt for Facebook’s army of fake likes

Come work with us!

I’m very pleased to announce that — along with George Danezis and Tomaso Aste, head of our Financial Computing group — I’ve been awarded a grant to continue our work on distributed ledgers (aka “blockchain-like things”) for the next three years.

Our group has already done a lot of research in this space, including George’s and my recent paper on centrally banked cryptocurrencies (at NDSS 2016) and Jens’ paper (along with Markulf Kohlweiss, a frequent UCL collaborator) on efficient ring signatures and applications to Zerocoin-style cryptocurrencies (at Eurocrypt 2015).  It’s great to have this opportunity to further investigate the challenges in this space and develop our vision for the future of these technologies, so big thanks to the EPSRC!

Anyway, the point of this post is to advertise, as part of this grant, three positions for postdoctoral researchers.  We are also seeking collaboration with any industrial partners investigating the potential usage of distributed ledgers, and in particular ones looking at the application of these ledgers across the following settings (or with a whole new setting in mind!):

  • Identity management. How can identities be stored, shared, and issued in a way that preserves privacy, prevents theft and fraud, and allows for informal forms of identity in places where no formal ones exist?
  • Supply chain transparency. How can supply chain information be stored in a way that proves integrity, preserves the privacy of individual actors, and can be presented to the end customer in a productive way?
  • Financial settlement. How can banking information be stored in a way that allows banks to easily perform gross settlement, reduces the burden on a central bank, and enables auditability of the proper functioning of the system?
  • Administration of benefits. How can benefits be administered to and used by disadvantaged populations in a way that preserves privacy, provides useful visibility into their spending, and protects against potential abuses of the system?

We expect the postdoctoral researchers to work with us and with each other on the many exciting problems in this space, which are spread across cryptography, computer and network security, behavioural economics, distributed systems, usable security, human-computer interaction, and software engineering (just to name a few!).  I encourage anyone interested to reach out to me (Sarah) to discuss this further, whether or not they’ve already done research on the particular topic of distributed ledgers.

That’s all for now, but please get in touch with me if you have any questions, and in the years to come I hope to invite many people to come work with us in London and to announce the various outcomes of this exciting project!

Bitcoin workshop at Financial Crypto 2016

On 26 February 2016 the 3rd workshop on Bitcoin and Blockchain Research in association with Financial Cryptography 2016 took place in Barbados. This workshop aims to bring together researchers interested in cryptocurrencies to present their latest work and discuss together the future of Bitcoin. The program chairs were Sarah Meiklejohn from University College London and Jeremy Clark from Concordia University. The themes addressed during the workshop included blockchain architecture, anonymity, and proof of work alternatives. This event was also a great way for researchers with similar interests to network and share their ideas.

The workshop consisted of 2 keynotes and 4 plenary sessions: Bitcoin network analysis, Enhancing Bitcoin, Ethereum, and Blockchain Architecture.

Nathaniel Popper kicked off the day with a keynote presentation. Nathaniel is a journalist from the New York Times and author of the book ‘Digital Gold: The Untold story of Bitcoin’. He went on to speak about the history of Bitcoin covering Silk Road, Mt Gox, as well as the role of governments.

Then the first session, about Bitcoin network analysis, included two talks. The first one, Stressing Out: Bitcoin Stress Testing, by Khaled Baqer et al., was about DoS attack on Bitcoin, and was presented by Ross Anderson due to visa issues. The second one was Why buy when you can rent? Bribery attacks on Bitcoin-style consensus, by Joseph Bonneau on bribery attacks and cloud mining.

The next session, Enhancing Bitcoin, started with a talk by Ethan Heilman, Blindly Signed Contracts: Anonymous On-Blockchain and Off-Blockchain Bitcoin Transactions, on how to enhance Bitcoin anonymity. Then Mathieu Turuani gave a talk on Automated Verification of Electrum wallet, followed by Aggelos Kiayias on Proof of Proof of Work. Today many light-weight clients use SPV verification instead of full verification. Is it possible to have an even lighter verification? They introduce a modification of the Bitcoin blockchain protocol with sublinear complexity in the length of the chain.

Continue reading Bitcoin workshop at Financial Crypto 2016

Biometrics for payments

HSBC and First Direct recently announced that they are introducing fingerprint and voice recognition authentication for customers of online and telephone banking. In my own research, I first found nearly 20 years ago that people who have a multitude of passwords and PINs cannot manage them as security experts want them to. As the number of digital devices and services we use has increased rapidly, managing dozens of login details has become a headache for most people. We recently reported that most bank customers juggle multiple PINs, and are unable to follow the rules that banks set in their contracts. Our research also found that many people dislike the 2-factor token solutions that are currently used by many UK banks.

Passwords as most people use them today are not particularly secure. Attackers can easily attempt to collect information on individuals, using leaks of password files not properly protected by some websites, “phishing” scams or malware planted on people’s computers. Reusing a banking password on other websites – something that many of us do because we cannot remember dozens of different passwords – is also a significant security risk.

The introduction of fingerprint recognition on smartphones – such as the iPhone – has delighted many users fed up with entering their PINs dozens of times a day. So the announcement that HSBC and other banks will be able to use the fingerprint sensor on their smartphones for banking means that millions of consumers will finally be able to end their battle with passwords and PINs and use biometrics instead. Other services people access from their smartphones are likely to follow suit. And given the negative impact that cumbersome authentication via passwords and PINs has on staff productivity and morale in many organisations, we can expect to see biometrics deployed in work contexts, too.

But while biometrics – unlike passwords – do not require mental gymnastics from users, there are different usability challenges. Leveraging the biometric from the modality of interaction – e.g. voice recognition phone-based interactions – makes authentication an easy task, but it will work considerably better in quiet environments than noisy ones – such as a train stations or with many people talking in the background. As many smartphone users have learnt, fingerprint sensors have a hard time recognising cold and wet fingers. And – as we report in a paper presented at IEEE Identity, Security and Behavior Analysis last week – privacy concerns mean some users ‘don’t like putting their face on the Internet’. Biometrics can’t come soon enough for most users, but there is still a lot of design and testing work to be done to make biometrics work for different interaction, physical and social contexts.

Privately gathering statistics and training simple models

Last week, Luca Melis has presented our NDSS16 paper “Efficient Private Statistics with Succinct Sketches“, where we show how to privately and efficiently aggregate data from many sources and/or large streams, and then use the aggregate to extract useful statistics and train simple machine learning models.

Our work is motivated by a few “real-world” problems:

  • Media broadcasting providers like the BBC (with which we collaborate) routinely collect data from their users about videos they have watched (e.g., on BBC’s iPlayer) in order to provide users with personalized suggestions for other videos, based on recommender systems like Item k-Nearest Neighbor (ItemKNN)
  • Urban and transport planning committees, such as London’s mass transport operators, need to gather statistics about people’s movements and commutes, e.g., to improve transportation services and predict near-future trends and anomalies on a short notice.
  • Network infrastructures like the Tor network need to gather traffic statistics, like the number of, and traffic generated by, Tor hidden services, in order to tune design decisions as well as convince their founders the infrastructure is used for the intended purposes.

While different in their application, these examples exhibit a common feature: the need for providers to aggregate large amounts of sensitive information from large numbers of data sources, in order to produce aggregate statistics and possibly train machine learning models.

Prior work has proposed a few cryptographic tools for privacy-enhanced computation that could be use for private collection of statistics. For instance, by relying on homomorphic encryption and/or secret sharing, an untrusted aggregator can receive encrypted readings from users and only decrypt their sum. However, these require users to perform a number of cryptographic operations, and transmit a number of ciphertexts, linear in the size of their inputs, which makes it impractical for the scenarios discussed above, whereby inputs to be aggregated are quite large. For instance, if we use ItemKNN for the recommendations, we would need to aggregate values for “co-views” (i.e., videos that have been watched by the same user) of hundreds of videos at the time – thus, each user would have to encrypt and transfer hundreds of thousands of values at the time.

Scaling private aggregation

We tackle the problem from two points of view: an “algorithmic” one and a “system” one. That is, we have worked both on the design of the necessary cryptographic and data structure tools, as well as on making it easy for application developers to easily support these tools in web and mobile applications.

Our intuition is that, in many scenarios, it might be enough to collect estimates of statistics and trade off an upper-bounded error with significant efficiency gains. For instance, the accuracy of a recommender system might not be really affected if the statistics we need to train the model are approximated with a small error.

Continue reading Privately gathering statistics and training simple models

“Do you see what I see?” ask Tor users, as a large number of websites reject them but accept non-Tor users

If you use an anonymity network such as Tor on a regular basis, you are probably familiar with various annoyances in your web browsing experience, ranging from pages saying “Access denied” to having to solve CAPTCHAs before continuing. Interestingly, these hurdles disappear if the same website is accessed without Tor. The growing trend of websites extending this kind of “differential treatment” to anonymous users undermines Tor’s overall utility, and adds a new dimension to the traditional threats to Tor (attacks on user privacy, or governments blocking access to Tor). There is plenty of anecdotal evidence about Tor users experiencing difficulties in browsing the web, for example the user-reported catalog of services blocking Tor. However, we don’t have sufficient detail about the problem to answer deeper questions like: how prevalent is differential treatment of Tor on the web; are there any centralized players with Tor-unfriendly policies that have a magnified effect on the browsing experience of Tor users; can we identify patterns in where these Tor-unfriendly websites are hosted (or located), and so forth.

Today we present our paper on this topic: “Do You See What I See? Differential Treatment of Anonymous Users” at the Network and Distributed System Security Symposium (NDSS). Together with researchers from the University of Cambridge, University College London, University of California, Berkeley and International Computer Science Institute (Berkeley), we conducted comprehensive network measurements to shed light on websites that block Tor. At the network layer, we scanned the entire IPv4 address space on port 80 from Tor exit nodes. At the application layer, we fetch the homepage from the most popular 1,000 websites (according to Alexa) from all Tor exit nodes. We compare these measurements with a baseline from non-Tor control measurements, and uncover significant evidence of Tor blocking. We estimate that at least 1.3 million IP addresses that would otherwise allow a TCP handshake on port 80 block the handshake if it originates from a Tor exit node. We also show that at least 3.67% of the most popular 1,000 websites block Tor users at the application layer.

Continue reading “Do you see what I see?” ask Tor users, as a large number of websites reject them but accept non-Tor users