PRIVACY versus UTILITY and CONTROL A False Dichotomy? Giuseppe Bianchi Roma, August 19, 2008 YICGG 2008 The motivating questions • Ubiquitous “any time, any place” e-presence • Pervasive services in a virtual pervasive world • Body networks, sensor networks, internet of things… Is pervasive technology forcing users to waive their privacy rights? And how much of our privacy can be traded for security and/or new service opportunities? Privacy issues in modern pervasive environments Tons of e-data gathered from… • • • • • …and processed: correlation, profiling, data mining, … • • • • • Pervasive always-on devices Body-connected devices/tags Sensing (smart) environments Biometrics E-surveillance Location Access communication Application …. Result: (where I am) (who I am) (who I’m connecting to) (what am I doing) • precise definitions of the individual’s preferences • Personal information on consumers perceived as asset often subject matter of commercial transactions • Individual often do not consent or even are not aware! Achieving Privacy Privacy: what it should be • the right of the individual to control his personal information by posing specific obligations on the subjects that process his data. What it is: • No consent = no service • Consent loose specification on how data will be treated Regulation fails from being precise (in most cases) In practice: consent = I accept to lose my privacy Privacy: a difficult problem Opposite and conflicting interests User privacy anonymity, untraceability, … How to guarantee against criminals and abuses? Public Interest Private activities Service operation public authorities and bodies must achieve public interests (governmental, security, economic, tax purposes, etc.) How to provide effective service without full data disclosure? information vital for commercial and customer relationships, business activities and strategies, … High privacy levels high level of control No control on users DICHOTOMY! high security Hard to run services high service utility The current solution: trust and hope Trust that the providers and their employeers respect your privacy (and the applicable laws) Trust that the security measures set up by the provider are adequate Privacy vs utility/control: trade-offs set forth only by regulation When public interest prevails E.g.: EU directives • Article 8 (2) of the European Convention of Human Rights and Fundamental Freedoms (Dec. 2000) • i) national security, ii) public safety, iii) economic well-being of the country, iv) prevention of disorder or crime, v) protection of health or morals, vi) protection of the rights and freedoms of others • Article 13 of the EU Directive 95/46/EC (Nov. 1995) • (a) national security; (b) defense; (c) public security; (d) prevention, investigation, detection and prosecution of criminal offences, or of breaches of ethics for regulated professions; (e) important economic or financial interest of a State, including monetary, budgetary and taxation matters; (f) monitoring, inspection or regulatory function connected with the exercise of official authority, (g) the protection of the data subject or of the rights and freedoms of others. • Also, (under certain limits) for scientific research and statistic creation purposes What data protection laws say EC/95/46 European Data Protection Directive Clear general principles but not nearly translated in a detailed set of mandatory security measures … Data must be processed fairly and lawfully Data must be collected for specified, explicit and legitimate purposes and not further processed in a way incompatible with those purposes Data must be adequate, relevant and not excessive in relation to the purposes for which they are collected and/or further processed Data must be kept in a form which permits identification of data subjects for no longer than is necessary for the purposes for which the data were collected or for which they are further processed … Consistent in ALL regulations!! Way too generic terms (how to enforce them?) a “minimum” of data security measures that must be implemented “fair” processing of the data “appropriate, adequate, suitable” data security measures that should be adopted Why laws and trust may not be the right answer DB The issue of insiders • how to tightly control employeers and sysadm? Security of long-term stored data is costly • Data retention laws pose long term storage obligations (6 months to 5 years) Many recent privacy violation scandals • GSM wiretapping in Italy, Greece, etc • Abuse of lawful interception hooks • Profiling abuses from operators Towards technology-based solutions Four major directions: • Empower users with improved control on their Example 1 data access and disclosure • Rely as much as possible on pseudonymization Example 2 and anonymization Example 2 • Divide et impera: distribute data across multiple entities, and permit recombination only by authorities Example 3 • Permit (selective and in-purpose) operation on data only if specific technically enforceable conditions arise Example 1: RFID Contactless Privacy Manager Faked hair Artificial leg model #4456 (polyester) model #459382 Developed in the IST-DISCREET EU project ref: [deliverables @ www.discreet.org, 2007-2008] Bakunin letters 5 x $100 banknotes Serial #: 597387,389473… Lingerie model Penelope Bras(Size 4) CPM: special RFID tag/reader; Selectively jams every access attempt Permits access only to user-authorized tags User-friendly GUI (on cellular phone) Connected via Bluetooth to CPM Example 2: pseudonymization PKI /1 Traditional access: • Identity = authenticate the user • Identity = control if user is authorized • Identity = use as accounting label A rethinking: • Authentication: no identity necessary we only need to have a way to get back to the user if abuses occur… • Authorization: no identity necessary Showing a permission does not necessarily imply identity disclosure • Accounting: a pseudonym is sufficient We need a stable label (alias for the user) and payment mechanisms Key idea: • Separate pseudonym from access permission Pseudonym: a label provided by a public infrastructure, which may be reverted only by authorities Access permission: a cryptographic “blind” credential Example 2: pseudonymization PKI /2 Developed in the IST-DISCREET EU project ref: [ENCTS 197, 2008] Pseudonym disclousure SP Server IR Server SP Administrative Domain ph IR Administrative Domain as ph e2 e1 as Multiple servers Chosen by the user Pseudonym not assigned by SP At registration, pseudonym gets “blindly” authorized (SP sees real user) – non trivial crypto extension needed During access, i) pseudonym gets verified, ii) authorization Is checked. No identity revealed Example 3: privacy-preserving network monitoring Developed in the IP packet (raw capture) Network Link Tap FRONT-END 1st stage (flow/feature) processing packet protection through encryption Protected packet BACK-END Packet meta-data analysis Escrowing (if suspected anomaly) Decryption and second-stage data processing IST-PRISM EU project ref: [ACM CCS NDA08, 2008] IDEA: allow decryption and processing of data ONLY IF specific conditions emerge Demonstrated on anomaly detection monitoring application. Ingredients: - Fast processing on front-end through Bloom filters - threshold-based decryption (Shamir’s approach) only when clearly formalized conditions representative of a suspected anomaly occur Conclusions Privacy: • Not only a regulatory issue • But a TECHNICAL issue! ICT can solve the (false!) privacy dichotomy! • Technical solutions do exist to have BOTH privacy AND utility & control! • Services and technologies themselves should be technically structured so as to comply with data protection Non ICT further research directions: • Law: how to quantitatively account for technical enhancements in next generation regulation? • Social/economic: can privacy preservation become an asset?