Nubian Rights Forum & 2 others v Attorney General & 6 others; Child Welfare Society & 9 others (Interested Parties) [2020] eKLR ‘identification’ is a process defined by a third party which confirms the identity of a person. In his view, NIIMS was an identification project as confirmed by statements made by the Respondents. 63. Mr. Anand stated that there was a possibility that people who had not enrolled would be rejected by the system as duplicates. He enumerated five concerns that he had in regard to NIIMS. The first was its purpose-free architecture, meaning that the data could be used for any purpose and has no limitation. On the other hand purposelimited architecture requires the data to be used only for specified purposes. 64. His second concern related to the enrolment system design. He stated that the data capture form asked for detailed information which is not required for giving a person a unique identity number that in other words, there is no data minimisation. Thirdly, he was concerned that the system was exclusionary by its architecture and design in that citizens may be prevented from getting their rights by, for instance, the requirement that they must have an existing identity card or good quality biometrics in order to be registered in NIIMS. 65. The fourth concern related to fear of mass surveillance. He expressed the view that government is able to conduct surveillance of its citizens from the data it collects because it has the decryption key. Further, that such surveillance is also a possibility with respect to functional and subsidiary databases which are not encrypted. In addition, that commercial surveillance can also arise from profiling of persons using the data collected. 66. Mr. Anand’s fifth concern related to breaches of the system and fraud. He stated that even though the Respondents alleged that the data would be protected through encryption, data must be decrypted before use, leading to security risks. He stated that hacking occurs when data is used and the encrypted data in this respect will need to be used for authentication purposes thus raising the opportunity for hacking. Mr. Anand explained that fraud is committed for a specific purpose, mostly financial gain, and centralising data raises the benefits of hacking to the fraudsters. 67. Finally, he expressed the opinion that the NIIMS system is archaic and there was failure to consider modern designs and alternatives. He gave examples of alternatives as including removing the linkages to functional databases such as in Austria and Estonia, or the use of PIN-based cards to authenticate identity instead of biometrics. 68. Questioned about the use and implication of open versus closed source technology, Mr. Anand stated that an open source technology project is constantly being peer-reviewed by a larger community of digital experts, which makes it more robust. With respect to whether such a system will deter terrorism, he stated that because of the large number of people trying to fix any loophole in the system, there is disclosure about loopholes which is a way of preventing the loopholes being used. He stated that in relation to NIIMS, there is no documentation or knowledge on the design, and one could not therefore know what loopholes there were in NIIMS. He stressed that in India, the information on the design of Aadhaar was in the public domain and he had even identified the loopholes and informed the government about it. 69. Mr. Anand stated that a cost-benefit analysis ought to have been carried out prior to implementation. He reiterated that NIIMS, being a digital infrastructure project, required a detailed cost-benefit analysis, and there was no evidence of such a process in NIIMS, which he concluded was not a well-thought out project. 70. Answering questions put to him in re-examination by Mr. Awele for the 2nd Petitioner, Mr. Anand stated that NIIMS is not deterministic, and that for a system to be deterministic, one needed to collect a lot of personal data using a multi-modal identity system. He stressed that the State will have to have a lot of data points for the system to be deterministic. He acknowledged that the Respondents had indicated in their replies their preference for closed source software. He had not, however, seen the design and architecture of the NIIMS system, but his view was that the government should have disclosed information about its technical architecture and algorithms to illustrate the http://www.kenyalaw.org - Page 12/174