1 1 00:00:00,330 --> 00:00:01,950 Data sharing. 2 2 00:00:01,950 --> 00:00:03,480 In this lesson, we're going to talk 3 3 00:00:03,480 --> 00:00:06,720 about data sharing and some of the agreements you might use. 4 4 00:00:06,720 --> 00:00:08,070 Now, one of the things I've mentioned earlier 5 5 00:00:08,070 --> 00:00:10,230 in this class is that you can outsource a service 6 6 00:00:10,230 --> 00:00:11,160 or an activity 7 7 00:00:11,160 --> 00:00:14,460 but you can't outsource the legal responsibility for it. 8 8 00:00:14,460 --> 00:00:15,870 Even if I give that job 9 9 00:00:15,870 --> 00:00:19,320 to somebody else to do for me, I'm still responsible for it. 10 10 00:00:19,320 --> 00:00:21,330 So, I can have somebody else do my bookkeeping 11 11 00:00:21,330 --> 00:00:22,290 but at the end of the year, 12 12 00:00:22,290 --> 00:00:24,930 the IRS is still going to come after me if it's done wrong. 13 13 00:00:24,930 --> 00:00:26,460 And so, these are the things you have to think about 14 14 00:00:26,460 --> 00:00:27,870 when you're dealing with data sharing 15 15 00:00:27,870 --> 00:00:30,660 because this kind of stuff can come back to bite you. 16 16 00:00:30,660 --> 00:00:31,530 Now we're going to talk about 17 17 00:00:31,530 --> 00:00:33,150 four specific types of agreements. 18 18 00:00:33,150 --> 00:00:36,270 We're going to talk about service level agreements or SLAs, 19 19 00:00:36,270 --> 00:00:38,641 interconnection security agreements or ISAs, 20 20 00:00:38,641 --> 00:00:41,130 non-disclosure agreements or NDAs, 21 21 00:00:41,130 --> 00:00:43,290 and data sharing and use agreements. 22 22 00:00:43,290 --> 00:00:44,201 Let's talk about each of these. 23 23 00:00:44,201 --> 00:00:46,020 Now, a service level agreement 24 24 00:00:46,020 --> 00:00:48,150 or an SLA is a contractual agreement. 25 25 00:00:48,150 --> 00:00:49,590 Setting out the detailed terms 26 26 00:00:49,590 --> 00:00:52,050 under which a service is being provided. 27 27 00:00:52,050 --> 00:00:54,090 So I might have an SLA 28 28 00:00:54,090 --> 00:00:57,360 between me and my ISP, my internet service provider. 29 29 00:00:57,360 --> 00:01:00,000 It says they're going to provide me X amount of storage space, 30 30 00:01:00,000 --> 00:01:02,431 X amount of bandwidth, X amount of uptime, 31 31 00:01:02,431 --> 00:01:05,040 and X amount of hours a day of support. 32 32 00:01:05,040 --> 00:01:07,200 That is all in contract, in writing 33 33 00:01:07,200 --> 00:01:09,630 that we have both signed and we both agreed to. 34 34 00:01:09,630 --> 00:01:11,250 Now, that means, I don't have to worry 35 35 00:01:11,250 --> 00:01:13,350 about hosting my website, right? 36 36 00:01:13,350 --> 00:01:14,183 Correct. 37 37 00:01:14,183 --> 00:01:15,630 But if the site gets hacked, 38 38 00:01:15,630 --> 00:01:17,400 guess who's still going to have to answer for it? 39 39 00:01:17,400 --> 00:01:19,620 Me, because I'm still responsible. 40 40 00:01:19,620 --> 00:01:21,300 My company is still responsible. 41 41 00:01:21,300 --> 00:01:23,941 We've outsourced the task of running the web server 42 42 00:01:23,941 --> 00:01:26,250 but we have not outsourced the responsibility. 43 43 00:01:26,250 --> 00:01:28,410 You can't give away that responsibility. 44 44 00:01:28,410 --> 00:01:29,580 The second one we're going to talk about 45 45 00:01:29,580 --> 00:01:32,760 is an interconnection security agreement or ISA. 46 46 00:01:32,760 --> 00:01:35,550 Now, an ISA is an agreement used by federal agencies. 47 47 00:01:35,550 --> 00:01:37,410 So, we're talking about the US government here 48 48 00:01:37,410 --> 00:01:39,279 to set out security risk awareness processes 49 49 00:01:39,279 --> 00:01:40,980 and commit the agency 50 50 00:01:40,980 --> 00:01:43,740 and suppliers to implementing security controls. 51 51 00:01:43,740 --> 00:01:44,573 Now, essentially, 52 52 00:01:44,573 --> 00:01:46,410 these are used when the federal government wants 53 53 00:01:46,410 --> 00:01:49,290 to allow a third-party to connect into their system. 54 54 00:01:49,290 --> 00:01:51,450 So, let's say, I got a big contract to take over 55 55 00:01:51,450 --> 00:01:54,090 all of the IT training for the federal government. 56 56 00:01:54,090 --> 00:01:55,020 They would want to connect 57 57 00:01:55,020 --> 00:01:57,750 into my web server to allow all their users access. 58 58 00:01:57,750 --> 00:01:58,590 We would have to have one of these 59 59 00:01:58,590 --> 00:02:00,120 interconnection security agreements 60 60 00:02:00,120 --> 00:02:00,953 and it might say 61 61 00:02:00,953 --> 00:02:03,120 that every three months, we have to do a vulnerability scan 62 62 00:02:03,120 --> 00:02:04,410 and give them a report. 63 63 00:02:04,410 --> 00:02:06,690 We have to have all patches done within 30 days. 64 64 00:02:06,690 --> 00:02:08,460 We have to have X, Y, and Z done, 65 65 00:02:08,460 --> 00:02:10,320 whatever those requirements are to make sure 66 66 00:02:10,320 --> 00:02:11,550 that our security requirements 67 67 00:02:11,550 --> 00:02:13,440 meet their level of agreement. 68 68 00:02:13,440 --> 00:02:15,570 That way, we're not just plugging bad networks 69 69 00:02:15,570 --> 00:02:17,070 into the federal government's network. 70 70 00:02:17,070 --> 00:02:18,870 That's what we're talking about with an ISA. 71 71 00:02:18,870 --> 00:02:21,330 Next, we have our third one, which is an NDA 72 72 00:02:21,330 --> 00:02:23,190 A non-disclosure agreement. 73 73 00:02:23,190 --> 00:02:25,680 This is a contract that sets forth the legal basis 74 74 00:02:25,680 --> 00:02:28,950 for protecting information assets between two parties. 75 75 00:02:28,950 --> 00:02:31,200 So, I'm making a new course, for instance, 76 76 00:02:31,200 --> 00:02:32,460 for another company. 77 77 00:02:32,460 --> 00:02:35,700 We have an NDA that says, I can't disclose 78 78 00:02:35,700 --> 00:02:38,760 what's in that course until it goes live and is published 79 79 00:02:38,760 --> 00:02:40,850 because they want to make sure that their competitors 80 80 00:02:40,850 --> 00:02:42,780 don't figure out what our table of contents 81 81 00:02:42,780 --> 00:02:44,940 is going to look like and build a competing course. 82 82 00:02:44,940 --> 00:02:46,110 That would be an NDA. 83 83 00:02:46,110 --> 00:02:48,120 We can't tell other people outside of the two 84 84 00:02:48,120 --> 00:02:49,250 of us what's going on. 85 85 00:02:49,250 --> 00:02:51,270 I have some other partners that I work with 86 86 00:02:51,270 --> 00:02:52,920 inside the business, and we have NDAs 87 87 00:02:52,920 --> 00:02:54,510 with those organizations as well. 88 88 00:02:54,510 --> 00:02:57,120 So, they can tell us things about future upcoming products 89 89 00:02:57,120 --> 00:02:59,670 they might be working on, but I can't tell you about it 90 90 00:02:59,670 --> 00:03:01,830 because you're not part of that NDA. 91 91 00:03:01,830 --> 00:03:03,000 Only the two of us are. 92 92 00:03:03,000 --> 00:03:04,800 And so we have this, basically it says, 93 93 00:03:04,800 --> 00:03:06,900 I'm going to keep your secrets and you're going to keep mine 94 94 00:03:06,900 --> 00:03:08,550 and neither of us will tell other people. 95 95 00:03:08,550 --> 00:03:10,009 That's what an NDA does. 96 96 00:03:10,009 --> 00:03:12,480 And then we have data sharing and use agreements. 97 97 00:03:12,480 --> 00:03:14,380 Now, these are an agreement that sets forth the terms 98 98 00:03:14,380 --> 00:03:17,550 under which personal data can be shared or used. 99 99 00:03:17,550 --> 00:03:19,140 So, let's say, you bought this course 100 100 00:03:19,140 --> 00:03:20,970 on some kind of a platform, and in part 101 101 00:03:20,970 --> 00:03:23,820 of their privacy policy, it says, they can monitor your use 102 102 00:03:23,820 --> 00:03:26,610 of the course, and they can use that to then sell that 103 103 00:03:26,610 --> 00:03:28,020 to third parties to tell them 104 104 00:03:28,020 --> 00:03:29,652 about e-learners in the market. 105 105 00:03:29,652 --> 00:03:31,890 They can then say, "Well, out of all the students, 106 106 00:03:31,890 --> 00:03:34,300 we have 50% finish all of their courses, 107 107 00:03:34,300 --> 00:03:36,660 20% by a second course. 108 108 00:03:36,660 --> 00:03:38,160 70% take courses in English. 109 109 00:03:38,160 --> 00:03:39,480 Whatever those facts and figures are, 110 110 00:03:39,480 --> 00:03:40,710 I'm just making that up. 111 111 00:03:40,710 --> 00:03:42,780 But they can share that data with a third-party 112 112 00:03:42,780 --> 00:03:45,150 if they have permission when they collected it from you 113 113 00:03:45,150 --> 00:03:48,030 and they now have a data sharing agreement in place. 114 114 00:03:48,030 --> 00:03:49,500 This is the idea of data sharing agreements. 115 115 00:03:49,500 --> 00:03:51,750 I can take my corporate data that I've collected off 116 116 00:03:51,750 --> 00:03:54,270 of those people that are using my service and then share it 117 117 00:03:54,270 --> 00:03:56,598 with somebody else under the terms of that agreement. 118 118 00:03:56,598 --> 00:03:59,310 Now, all of these things predicate themselves 119 119 00:03:59,310 --> 00:04:02,040 on us having personal data on our users 120 120 00:04:02,040 --> 00:04:04,560 and this becomes a large data set for us. 121 121 00:04:04,560 --> 00:04:07,980 Now, these data sets may be subject to pseudonymization 122 122 00:04:07,980 --> 00:04:10,890 or de-identification to remove personal data. 123 123 00:04:10,890 --> 00:04:13,650 Now, we've talked a lot about privacy inside this course 124 124 00:04:13,650 --> 00:04:14,940 and one of the ideas here 125 125 00:04:14,940 --> 00:04:17,190 is that we can actually take your data 126 126 00:04:17,190 --> 00:04:19,560 and we can de-identify you from it. 127 127 00:04:19,560 --> 00:04:22,140 And that makes that data no longer personal data. 128 128 00:04:22,140 --> 00:04:23,940 It just becomes big data. 129 129 00:04:23,940 --> 00:04:26,760 And this is done a lot by a lot of different companies, 130 130 00:04:26,760 --> 00:04:29,278 the Googles and the Facebooks and others of the world. 131 131 00:04:29,278 --> 00:04:31,590 Even if they don't have permission, to have 132 132 00:04:31,590 --> 00:04:33,510 your personal data, because you have opted out, 133 133 00:04:33,510 --> 00:04:35,970 there might be information about you 134 134 00:04:35,970 --> 00:04:39,390 at large that could be used if they de-identify you. 135 135 00:04:39,390 --> 00:04:41,520 And what this means is we want to take out anything 136 136 00:04:41,520 --> 00:04:43,265 that can identify you uniquely. 137 137 00:04:43,265 --> 00:04:45,660 So, let's say, I'm an online training company 138 138 00:04:45,660 --> 00:04:47,657 and I have lots of students in my site 139 139 00:04:47,657 --> 00:04:51,210 but I never asked you if I could share your personal data. 140 140 00:04:51,210 --> 00:04:53,340 Well, that means, I can't share your personal data 141 141 00:04:53,340 --> 00:04:56,790 but I can de-identify you and then use that data. 142 142 00:04:56,790 --> 00:04:58,015 And what I mean by this is I could say, 143 143 00:04:58,015 --> 00:05:00,495 out of the 100,000 students who took this course, 144 144 00:05:00,495 --> 00:05:02,804 99% passed on their first try. 145 145 00:05:02,804 --> 00:05:04,470 Now that doesn't identify you 146 146 00:05:04,470 --> 00:05:06,000 as one of my students in any way. 147 147 00:05:06,000 --> 00:05:08,430 It doesn't identify you in the least. 148 148 00:05:08,430 --> 00:05:10,170 You are just one of those 100,000. 149 149 00:05:10,170 --> 00:05:11,850 So, that is now de-identified. 150 150 00:05:11,850 --> 00:05:15,090 Or I can take smaller groups of information from my course. 151 151 00:05:15,090 --> 00:05:18,870 For instance, I could say males who are 35 to 40 years old 152 152 00:05:18,870 --> 00:05:21,300 pass at a rate of 80% or higher. 153 153 00:05:21,300 --> 00:05:25,800 Females who are 18 to 24 pass at 99%. 154 154 00:05:25,800 --> 00:05:26,970 And again, I'm just making up numbers. 155 155 00:05:26,970 --> 00:05:28,680 This doesn't, these aren't real numbers 156 156 00:05:28,680 --> 00:05:30,810 but it gives you the idea of how we can take subsets 157 157 00:05:30,810 --> 00:05:33,960 of the larger data and it doesn't identify you uniquely. 158 158 00:05:33,960 --> 00:05:34,793 This is the idea 159 159 00:05:34,793 --> 00:05:37,860 of de-identification or pseudonymization, 160 160 00:05:37,860 --> 00:05:40,200 by taking away those personal details and making it 161 161 00:05:40,200 --> 00:05:42,600 about the larger group instead of an individual.