1 00:00:01,400 --> 00:00:07,160 Speaker 1: It's Night Side with Dan Ray on WVZ, Boston's news radio. 2 00:00:07,760 --> 00:00:09,840 Speaker 2: Well, this is an interesting story that was brought to 3 00:00:09,880 --> 00:00:12,920 Speaker 2: my attention earlier this week, and I thought it's something 4 00:00:12,960 --> 00:00:15,240 Speaker 2: that you should be aware of if you're a customer 5 00:00:15,360 --> 00:00:18,960 Speaker 2: at Wegmans, particularly if you're a customer in New York City. 6 00:00:19,560 --> 00:00:23,560 Speaker 2: Apparently Wegmans has decided I shouldn't say, Apparently they have 7 00:00:23,680 --> 00:00:29,840 Speaker 2: admitted that they decided to use some facial recognition technology. 8 00:00:30,760 --> 00:00:34,080 Speaker 2: Just read you briefly from their statement if I could. 9 00:00:34,680 --> 00:00:37,080 Speaker 2: At Wegmans, the safety of our customers and employees is 10 00:00:37,159 --> 00:00:39,720 Speaker 2: a top priority. Like many retailers, we use cameras to 11 00:00:39,760 --> 00:00:43,040 Speaker 2: help identify individuals who post a risk to our people, 12 00:00:43,080 --> 00:00:46,519 Speaker 2: customers or operation. In a small fraction of our stores 13 00:00:46,600 --> 00:00:50,640 Speaker 2: located in communities that exhibit an elevated risk, we're deployed 14 00:00:50,640 --> 00:00:54,840 Speaker 2: cameras equipped with facial recognition technology in New York City. 15 00:00:55,400 --> 00:00:58,880 Speaker 2: We comply with local requirements by posting the mandated signage 16 00:00:58,880 --> 00:01:03,000 Speaker 2: to notify customers about the technology. The technology is solely 17 00:01:03,120 --> 00:01:05,959 Speaker 2: used for keeping our stores secure and safe. The system 18 00:01:06,000 --> 00:01:09,280 Speaker 2: collects facial recognition data and only uses it to identify 19 00:01:09,280 --> 00:01:13,240 Speaker 2: individuals who have been previously flagged for misconduct. We do 20 00:01:13,319 --> 00:01:17,319 Speaker 2: not collect biometric data such as retinal scans or voice prints. 21 00:01:17,360 --> 00:01:20,759 Speaker 2: Images and video are retained only for as long as 22 00:01:20,840 --> 00:01:24,959 Speaker 2: necessary for security purposes, and then they are disposed of 23 00:01:25,240 --> 00:01:28,160 Speaker 2: for security reason. We do not disclose the exact retension period, 24 00:01:28,160 --> 00:01:32,360 Speaker 2: but it aligns with industry standards. Bottom line is they say, 25 00:01:32,440 --> 00:01:35,520 Speaker 2: our goal is simple to keep our stores safe and secure. Now, 26 00:01:35,560 --> 00:01:37,800 Speaker 2: I am not a customer of Wegmans. I know that 27 00:01:37,800 --> 00:01:40,800 Speaker 2: there are Wegmans in the Boston area and throughout our 28 00:01:40,840 --> 00:01:46,319 Speaker 2: listening audience. And to discuss this tonight is Professor Eric 29 00:01:46,520 --> 00:01:51,040 Speaker 2: Learned Miller. Professor Miller is a professor and chair of 30 00:01:51,080 --> 00:01:54,160 Speaker 2: the faculty at the UMAs Amherst Manning College of Information 31 00:01:54,240 --> 00:01:59,200 Speaker 2: and Computer Sciences. He knows a lot about facial recognition. 32 00:01:59,240 --> 00:02:02,440 Speaker 2: I spoke with him today And you have been involved 33 00:02:03,000 --> 00:02:07,760 Speaker 2: with facial recognition, Professor Learned Miller for how many years? 34 00:02:09,520 --> 00:02:12,320 Speaker 3: Hi? First of all, thanks for inviting me, Thank you 35 00:02:12,320 --> 00:02:16,720 Speaker 3: for joining us. Yeah, I've been doing this, you know, 36 00:02:16,840 --> 00:02:20,200 Speaker 3: computer vision more generally since about nineteen ninety five, and 37 00:02:20,720 --> 00:02:24,400 Speaker 3: focused on face recognition starting in around two thousand and five. 38 00:02:24,880 --> 00:02:29,040 Speaker 2: So yeah, you have been in on the ground floor 39 00:02:29,240 --> 00:02:36,639 Speaker 2: of this And like any you know, electronic advancement, scientific development, 40 00:02:36,680 --> 00:02:40,040 Speaker 2: whatever you want to characterize it, it certainly has its benefits, 41 00:02:40,440 --> 00:02:45,240 Speaker 2: but it has its its potential abuses and detriments. You've 42 00:02:45,280 --> 00:02:49,160 Speaker 2: read this story. You were not involved in Wegman's decision 43 00:02:49,240 --> 00:02:51,200 Speaker 2: to do this. I want to separate you from that 44 00:02:51,240 --> 00:02:53,239 Speaker 2: with the audience and make sure that you're not here 45 00:02:53,560 --> 00:02:56,440 Speaker 2: as a representative in any way, shape or form of Wegman's. 46 00:02:56,480 --> 00:03:03,600 Speaker 2: But does this one surprise you? And what concerns, if any, 47 00:03:03,880 --> 00:03:08,760 Speaker 2: should people have about this development and disclosure? 48 00:03:10,160 --> 00:03:13,720 Speaker 3: Absolutely? Well, you know, first of all, I think there 49 00:03:13,720 --> 00:03:17,240 Speaker 3: are people out out there on both sides of this 50 00:03:17,400 --> 00:03:20,800 Speaker 3: issue that have the sort of the you know, the 51 00:03:20,840 --> 00:03:24,840 Speaker 3: strongest possible position, like we should ban all face recognition, 52 00:03:25,480 --> 00:03:28,240 Speaker 3: or we you know, or there shouldn't be any restrictions 53 00:03:28,280 --> 00:03:31,400 Speaker 3: on it. And and I'm not in either of those camps. 54 00:03:31,480 --> 00:03:35,960 Speaker 3: I realize it's useful for a lot of things. For example, 55 00:03:36,080 --> 00:03:39,000 Speaker 3: I've written papers with my father about using it to 56 00:03:39,120 --> 00:03:44,840 Speaker 3: diagnose disease that's been undiagnosed. And I was also on 57 00:03:44,880 --> 00:03:47,520 Speaker 3: a commission in the state of Massachusetts. I was the 58 00:03:48,160 --> 00:03:54,120 Speaker 3: Governor Baker's designee on his Commission to examine the use 59 00:03:54,160 --> 00:03:58,040 Speaker 3: of face recognition by police in the state of Massachusetts. 60 00:03:58,520 --> 00:04:01,640 Speaker 3: And I can talk more about that experience later, but 61 00:04:01,680 --> 00:04:07,040 Speaker 3: it was a really great cooperation between police officers who 62 00:04:07,120 --> 00:04:09,520 Speaker 3: want to get the benefit of it, and also civil 63 00:04:09,640 --> 00:04:13,960 Speaker 3: rights advocates and lawyers who want to make sure that 64 00:04:14,080 --> 00:04:17,600 Speaker 3: it's used appropriately with a proper precaution. So when I 65 00:04:17,640 --> 00:04:22,400 Speaker 3: see something like the Wegman's thing, you know, my reaction is, yeah, 66 00:04:22,480 --> 00:04:26,640 Speaker 3: that's a natural evolution of the companies that are trying 67 00:04:26,640 --> 00:04:32,280 Speaker 3: to sell this technology. They're trying to convince people that 68 00:04:32,360 --> 00:04:34,920 Speaker 3: it's going to help their bottom line, and it may 69 00:04:35,000 --> 00:04:41,160 Speaker 3: well do that. And my general position is that if 70 00:04:41,160 --> 00:04:44,599 Speaker 3: you have enough safeguards in place, then it can be 71 00:04:44,640 --> 00:04:48,760 Speaker 3: done safely, and we just have to protect people against 72 00:04:48,800 --> 00:04:52,720 Speaker 3: the cases when it will inevitably make errors and make 73 00:04:52,760 --> 00:04:54,960 Speaker 3: sure we can unwind those errors. 74 00:04:55,600 --> 00:04:58,600 Speaker 2: Here's what strikes me as odd about this, and I'd 75 00:04:58,640 --> 00:05:00,680 Speaker 2: love to know if you agree with me disagree with me. 76 00:05:00,800 --> 00:05:02,640 Speaker 2: You know more about it than I do, so feel 77 00:05:02,680 --> 00:05:07,719 Speaker 2: free to to explain to me why I'm mistaken. Is 78 00:05:07,760 --> 00:05:12,200 Speaker 2: that I can understand if you're running a high end 79 00:05:12,440 --> 00:05:16,400 Speaker 2: jewelry store, will you have security concerns and you have 80 00:05:16,520 --> 00:05:20,479 Speaker 2: people who are able to come in and basically boost 81 00:05:21,200 --> 00:05:25,120 Speaker 2: items that are quite expensive. But even in an upscale 82 00:05:25,200 --> 00:05:30,640 Speaker 2: grocery store like Wegmans, I just I think is this 83 00:05:30,760 --> 00:05:36,640 Speaker 2: an overreaction? And I wonder how customers of Wegmans will react, 84 00:05:36,680 --> 00:05:41,039 Speaker 2: and if they're so wedded to the store and the 85 00:05:41,120 --> 00:05:46,320 Speaker 2: allure of shopping at Wegmans, they will not worry about it. 86 00:05:46,400 --> 00:05:50,440 Speaker 2: But why do you think a grocery store, you know, again, 87 00:05:50,480 --> 00:05:56,400 Speaker 2: what is someone going to do walk out with, you know, 88 00:05:55,240 --> 00:06:01,560 Speaker 2: a package? I don't know. Yeah, I mean, it just 89 00:06:01,560 --> 00:06:04,719 Speaker 2: seems to me it's a little bit of an over reaction. 90 00:06:05,000 --> 00:06:08,040 Speaker 2: And you would think that if there are people who 91 00:06:08,080 --> 00:06:10,080 Speaker 2: are just problematic. I mean, if you live in an 92 00:06:10,120 --> 00:06:13,400 Speaker 2: area where people who are you know, drunk and disorderly 93 00:06:13,520 --> 00:06:17,320 Speaker 2: are coming into your store and using it for you know, 94 00:06:17,440 --> 00:06:20,960 Speaker 2: just to get warm or trying to use restrooms, that 95 00:06:20,960 --> 00:06:25,320 Speaker 2: that maybe you prefer the thing not. Whatever the concern is, 96 00:06:25,480 --> 00:06:29,600 Speaker 2: I kind of imagined that Wegmans is losing a lot 97 00:06:29,920 --> 00:06:35,120 Speaker 2: of stuff to theft. I'm trying to reconcile Wegman's and 98 00:06:35,240 --> 00:06:37,479 Speaker 2: facial recognition, and I'm having trouble doing that. 99 00:06:38,080 --> 00:06:40,880 Speaker 3: Well, you know. I mean that's I think you're bringing 100 00:06:40,960 --> 00:06:44,599 Speaker 3: up excellent marketing points, like how many customers are you 101 00:06:44,640 --> 00:06:47,239 Speaker 3: going to turn off? I mean some some may feel safer, 102 00:06:47,640 --> 00:06:52,200 Speaker 3: some may feel that less safe, or that they're being 103 00:06:53,360 --> 00:06:56,039 Speaker 3: washed and things like that. So you know, I'm certainly 104 00:06:56,080 --> 00:06:59,760 Speaker 3: not a grocery store chain manager and don't know how 105 00:06:59,800 --> 00:07:03,160 Speaker 3: to make those decisions, but I think they are definitely 106 00:07:03,200 --> 00:07:05,640 Speaker 3: people who are going to respond in different ways. One 107 00:07:05,640 --> 00:07:09,400 Speaker 3: thing that struck me about the story was initially I 108 00:07:09,480 --> 00:07:13,119 Speaker 3: assumed it was probably about shoplifting, right, like, let's make sure, 109 00:07:13,680 --> 00:07:18,920 Speaker 3: let's make sure that repeat shoplifters are that we can 110 00:07:18,960 --> 00:07:20,920 Speaker 3: tamp down on that, And I was a little bit 111 00:07:21,000 --> 00:07:25,000 Speaker 3: surprised to see them talk about safety as well. And 112 00:07:25,160 --> 00:07:29,840 Speaker 3: that makes me like, honestly, if there are safety incidents, 113 00:07:29,920 --> 00:07:32,960 Speaker 3: I would think you would want, you know, boots on 114 00:07:33,000 --> 00:07:36,520 Speaker 3: the ground, people, you know, you would want a because 115 00:07:36,560 --> 00:07:41,320 Speaker 3: if you recognize somebody on a camera and things have 116 00:07:41,360 --> 00:07:44,680 Speaker 3: already gotten out of hand, I don't I think you're 117 00:07:44,720 --> 00:07:49,600 Speaker 3: still going to need security. So I don't know, I'm it, 118 00:07:50,600 --> 00:07:54,040 Speaker 3: I wonder well, I don't want to speculate, but but 119 00:07:54,160 --> 00:07:58,080 Speaker 3: I think, you know, I was I was surprised to 120 00:07:58,120 --> 00:08:01,960 Speaker 3: see the security aspect, and thought it would be more about, 121 00:08:02,640 --> 00:08:06,520 Speaker 3: you know, uh, reducing theft and helping. 122 00:08:06,280 --> 00:08:07,080 Speaker 4: Their bottom line. 123 00:08:07,240 --> 00:08:08,960 Speaker 3: But but I really don't know. 124 00:08:09,320 --> 00:08:14,400 Speaker 2: Yeah, I have read some stories about shoplifting in supermarkets 125 00:08:14,840 --> 00:08:18,320 Speaker 2: that sometimes people come in with, particularly with the time, 126 00:08:18,400 --> 00:08:22,600 Speaker 2: with heavy overcoats, and you know, they'll try to put 127 00:08:22,760 --> 00:08:26,400 Speaker 2: you know, high quality cuts of beef inside those overcoats. 128 00:08:26,800 --> 00:08:27,240 Speaker 5: Uh. 129 00:08:27,280 --> 00:08:30,400 Speaker 2: And I could see that that that, but but it 130 00:08:30,520 --> 00:08:34,160 Speaker 2: seems to me that it's a bit of an overreaction. Uh. 131 00:08:34,360 --> 00:08:37,160 Speaker 2: And and we're going to invite callers to call with 132 00:08:37,320 --> 00:08:41,440 Speaker 2: questions and and also comments and observations. I'm not a 133 00:08:41,440 --> 00:08:44,200 Speaker 2: Wegman shopper, to be really honest with you. There's one 134 00:08:44,240 --> 00:08:47,880 Speaker 2: that's near me, but it's in a shopping center and 135 00:08:47,920 --> 00:08:50,960 Speaker 2: it's a it's awkward to get into, awkward to park. 136 00:08:52,320 --> 00:08:55,360 Speaker 2: It's and and the one time that I went in there, 137 00:08:55,520 --> 00:09:00,280 Speaker 2: I found it really cramped. And so I'm I'm I'm 138 00:09:00,280 --> 00:09:03,720 Speaker 2: not a Wegman's guy, but I'll be interested if there's 139 00:09:03,760 --> 00:09:06,280 Speaker 2: some Wegman shoppers who might be able to give us 140 00:09:06,280 --> 00:09:11,320 Speaker 2: some insights as well. We're talking with UMass Manning College 141 00:09:11,360 --> 00:09:15,520 Speaker 2: of Information and Computer Science is the chair of that department. 142 00:09:15,960 --> 00:09:18,360 Speaker 2: They had the chair of the faculty at UMass of 143 00:09:18,400 --> 00:09:22,760 Speaker 2: this department, Professor Eric Learned Miller, who has been involved 144 00:09:24,120 --> 00:09:29,160 Speaker 2: in the really the development of facial recognition technology since 145 00:09:30,160 --> 00:09:34,560 Speaker 2: uh late last century and early into this century. So 146 00:09:34,600 --> 00:09:37,280 Speaker 2: he's he knows more much more about that than I do. 147 00:09:37,960 --> 00:09:40,200 Speaker 2: But I also know that within my audience there are 148 00:09:40,200 --> 00:09:42,160 Speaker 2: people who have points of view. So if you would 149 00:09:42,160 --> 00:09:44,720 Speaker 2: like to join the conversation and you take a couple 150 00:09:44,800 --> 00:09:47,840 Speaker 2: of questions, I assume professor, if you'd be so kind, 151 00:09:48,240 --> 00:09:51,600 Speaker 2: that would be great. Thank you for doing that. Six 152 00:09:51,720 --> 00:09:56,000 Speaker 2: one seven two thirty six one seven, nine thirty is 153 00:09:56,040 --> 00:10:01,280 Speaker 2: the number. And really at the at the the baseline, well, uh, 154 00:10:01,520 --> 00:10:04,240 Speaker 2: cameras are everywhere. I think all of us know that, 155 00:10:04,280 --> 00:10:06,640 Speaker 2: but this is a little more than a camera. This 156 00:10:06,760 --> 00:10:10,760 Speaker 2: is facial recognition. Uh and and we'll we'll dive into 157 00:10:10,800 --> 00:10:12,760 Speaker 2: that a little more. But i'd love to know if 158 00:10:12,800 --> 00:10:16,880 Speaker 2: you found out that your supermarket that you go to regularly, 159 00:10:17,040 --> 00:10:23,640 Speaker 2: or your local bodega, for example, has facial recognition videos. 160 00:10:24,520 --> 00:10:26,960 Speaker 2: All of us understand that you that when you're on 161 00:10:27,040 --> 00:10:31,840 Speaker 2: public streets, you are going to be uh A photographed 162 00:10:32,080 --> 00:10:36,679 Speaker 2: there everybody, I think understands that. Now, however, when you're 163 00:10:36,720 --> 00:10:43,199 Speaker 2: going into a private supermarket, not that you would anticipate privacy, 164 00:10:43,320 --> 00:10:47,760 Speaker 2: but the fact that your your facial recognition, and it 165 00:10:47,760 --> 00:10:51,640 Speaker 2: looks as if, according to this statement from Wegman's, they 166 00:10:51,679 --> 00:10:54,760 Speaker 2: intend to store this stuff for a while, and they 167 00:10:54,800 --> 00:10:58,240 Speaker 2: did not disclose the exact retension period. When I get back, 168 00:10:58,280 --> 00:11:00,520 Speaker 2: I want to ask you, professor about that. Why they 169 00:11:00,559 --> 00:11:04,800 Speaker 2: would want to store it. Obviously, if they needed someone 170 00:11:05,600 --> 00:11:08,160 Speaker 2: who they felt was suspicious, I could see that, But 171 00:11:08,240 --> 00:11:13,880 Speaker 2: why would they suppose store the facial recognition of you know, 172 00:11:14,080 --> 00:11:16,320 Speaker 2: elderly folks who are just coming in and trying to 173 00:11:16,320 --> 00:11:20,280 Speaker 2: buy their groceries. Well, last I'll reask that question right 174 00:11:20,320 --> 00:11:24,200 Speaker 2: after this break. My guest again with us tonight from 175 00:11:24,280 --> 00:11:28,480 Speaker 2: U mass Amherst, the Chair of the Faculty of the Faculty, 176 00:11:28,559 --> 00:11:31,800 Speaker 2: Chair of the Chair of the Faculty at the UMSS 177 00:11:31,800 --> 00:11:35,920 Speaker 2: Amherst Manning College of Information and Computer Sciences, Doctor Eric 178 00:11:36,040 --> 00:11:39,959 Speaker 2: Learned Miller back on Nightside right after this, You're. 179 00:11:39,800 --> 00:11:47,400 Speaker 1: On Night Side with Dan Ray on Boston's News Radio. 180 00:11:46,360 --> 00:11:50,840 Speaker 2: With me from U mass Amherst, the Manning College of 181 00:11:50,880 --> 00:11:53,920 Speaker 2: Information and Computer Sciences, The Professor and Chair of the 182 00:11:53,960 --> 00:11:59,360 Speaker 2: faculty UH at that distinguished University U mass Amherst, and 183 00:11:59,400 --> 00:12:02,960 Speaker 2: we're talking with him. His name is as Professor Eric 184 00:12:03,160 --> 00:12:08,800 Speaker 2: Leernet Miller, Professor. In this statement from Wegman's they say 185 00:12:09,280 --> 00:12:12,920 Speaker 2: we do not disclose the exact retention period because of 186 00:12:13,000 --> 00:12:19,000 Speaker 2: security purposes. Why would they hold videos of an average 187 00:12:19,040 --> 00:12:22,760 Speaker 2: customer who goes in, buy some products and walks out 188 00:12:22,800 --> 00:12:25,040 Speaker 2: the door without incident. I can see why they might 189 00:12:25,120 --> 00:12:28,040 Speaker 2: hold the videos of people who have given them a 190 00:12:28,080 --> 00:12:31,760 Speaker 2: problem or cause them some concern. What's the benefit of 191 00:12:31,800 --> 00:12:37,000 Speaker 2: holding videos of customers who come in and do what 192 00:12:37,040 --> 00:12:39,079 Speaker 2: customer is supposed to do, buy some product and. 193 00:12:39,120 --> 00:12:43,480 Speaker 3: Leave good question? I mean, you know, this would just 194 00:12:43,520 --> 00:12:50,400 Speaker 3: be speculation, But you know, sometimes if they do see 195 00:12:50,440 --> 00:12:54,800 Speaker 3: an incident, they could possibly go back in time and 196 00:12:54,920 --> 00:12:59,440 Speaker 3: re examine old footage right that perhaps that they hadn't 197 00:12:59,440 --> 00:13:02,280 Speaker 3: looked at before and see if there's a pattern or 198 00:13:02,320 --> 00:13:05,959 Speaker 3: that kind of thing. So that would be one obvious reason. 199 00:13:06,880 --> 00:13:11,160 Speaker 3: I mean, they're trying to establish patterns, and that might 200 00:13:11,240 --> 00:13:14,640 Speaker 3: take you know, multiple views of the same person. I 201 00:13:14,679 --> 00:13:15,640 Speaker 3: guess we did. 202 00:13:17,400 --> 00:13:19,800 Speaker 2: We did, for the record reach out to Wegmans and 203 00:13:19,840 --> 00:13:22,800 Speaker 2: see if they would have someone to join us tonight 204 00:13:22,840 --> 00:13:27,560 Speaker 2: as well. Uh, is this something that in your experience, 205 00:13:27,600 --> 00:13:31,800 Speaker 2: your vast experience going back now, you know, many many years, 206 00:13:32,600 --> 00:13:36,240 Speaker 2: is this something that's would become more commonplace at you know, 207 00:13:36,320 --> 00:13:39,800 Speaker 2: again grocery stores. I just you know, if this was 208 00:13:40,280 --> 00:13:44,880 Speaker 2: you know, the high the highest end jewelry businesses or 209 00:13:45,040 --> 00:13:47,520 Speaker 2: or something like that where people do have the ability 210 00:13:47,559 --> 00:13:52,720 Speaker 2: to you know, to steal something of great value. Do 211 00:13:52,760 --> 00:13:54,960 Speaker 2: you think this is part of our brave new world 212 00:13:55,040 --> 00:13:55,520 Speaker 2: we're going to have? 213 00:13:56,280 --> 00:13:58,960 Speaker 3: Yeah, well, i'd say I'd say a couple of things. 214 00:13:59,000 --> 00:14:04,520 Speaker 3: I mean, there's steathly a slow well creek makes it 215 00:14:04,559 --> 00:14:08,440 Speaker 3: all sound negative, but there's definitely a slow increasing deployment 216 00:14:08,480 --> 00:14:12,440 Speaker 3: of this technology. For example, ten years ago, nobody was 217 00:14:12,520 --> 00:14:16,400 Speaker 3: using it at at checking gates at airports, and now 218 00:14:16,440 --> 00:14:20,000 Speaker 3: it's now it's quite regular in airport security. Not everywhere, 219 00:14:20,000 --> 00:14:22,920 Speaker 3: but in many places. So you see it more and 220 00:14:23,000 --> 00:14:28,040 Speaker 3: more and more places. My guess is, you know the 221 00:14:28,160 --> 00:14:31,880 Speaker 3: companies that developed the software, you know, they have sales 222 00:14:31,920 --> 00:14:34,760 Speaker 3: teams at Their job, of course, is to sell it 223 00:14:34,880 --> 00:14:37,920 Speaker 3: to as many people as they can. So their pitch 224 00:14:38,280 --> 00:14:40,600 Speaker 3: is going to be that this is going to save 225 00:14:40,600 --> 00:14:45,400 Speaker 3: your bottom line. So so for example, let's say Bregmans has. 226 00:14:45,720 --> 00:14:48,880 Speaker 3: Let's say they already had video surveillance and they had 227 00:14:48,960 --> 00:14:52,400 Speaker 3: maybe you know, two people sitting there watching their video 228 00:14:52,480 --> 00:14:58,200 Speaker 3: surveillance for activities or problems. And I can imagine that 229 00:14:59,000 --> 00:15:02,960 Speaker 3: a face recognition company might easily say something like, look, 230 00:15:03,280 --> 00:15:06,840 Speaker 3: you can easily pull it down to one or even 231 00:15:06,880 --> 00:15:11,400 Speaker 3: one part time person and we're going to replace for you, 232 00:15:11,400 --> 00:15:14,080 Speaker 3: you know, two, one or two or three of your employees, 233 00:15:14,480 --> 00:15:16,600 Speaker 3: and that's going to save you big on the bottom line. 234 00:15:16,640 --> 00:15:19,280 Speaker 3: You're going to save one hundred k every year by 235 00:15:19,400 --> 00:15:21,800 Speaker 3: not hiring those people, and your security is going to 236 00:15:21,800 --> 00:15:24,280 Speaker 3: be just as good, if not better. So you know, 237 00:15:24,480 --> 00:15:29,400 Speaker 3: there's there's always, almost always the companies that sell this 238 00:15:29,480 --> 00:15:33,440 Speaker 3: stuff are are are trying to convince you that you're 239 00:15:33,440 --> 00:15:36,600 Speaker 3: going to be able to do the same thing, possibly better, 240 00:15:37,040 --> 00:15:42,280 Speaker 3: for less cost. So that's my guess is that somebody 241 00:15:42,360 --> 00:15:46,760 Speaker 3: managed to convince the Leggings folks. Now they may not 242 00:15:46,880 --> 00:15:49,920 Speaker 3: have thought about what you brought up that some customers 243 00:15:51,160 --> 00:15:53,920 Speaker 3: might be turned off by it. And I don't know 244 00:15:53,920 --> 00:15:55,920 Speaker 3: if we're waiting for a caller, but I but I 245 00:15:55,960 --> 00:16:00,600 Speaker 3: do want to say that for this afternoon, I I 246 00:16:00,680 --> 00:16:07,000 Speaker 3: put your face into a publicly available face recognition search engine. Okay, 247 00:16:07,160 --> 00:16:13,160 Speaker 3: and that's an interesting results most can I go ahead 248 00:16:13,160 --> 00:16:13,440 Speaker 3: with this? 249 00:16:13,920 --> 00:16:16,360 Speaker 2: Let me ask you, is there anything about the cattle 250 00:16:16,480 --> 00:16:21,920 Speaker 2: rustling felony from Utah back in the nineteen twenties, because 251 00:16:21,920 --> 00:16:27,520 Speaker 2: that wasn't me only kidding. I trust you. Look, I'm 252 00:16:27,560 --> 00:16:30,440 Speaker 2: an open book, I'm a public figure. What do you got? 253 00:16:30,680 --> 00:16:32,440 Speaker 2: Go ahead hit me so. 254 00:16:33,640 --> 00:16:36,600 Speaker 3: Well, I will say that. You know, I got about 255 00:16:36,600 --> 00:16:40,120 Speaker 3: fifty pictures back, and forty nine of them were in 256 00:16:40,200 --> 00:16:44,160 Speaker 3: fact you. But there was one I was like, that's 257 00:16:44,200 --> 00:16:48,880 Speaker 3: definitely not Dan, and there was no name associated with 258 00:16:48,960 --> 00:16:51,800 Speaker 3: it unless I paid like a fifty dollars fee. So 259 00:16:51,960 --> 00:16:57,240 Speaker 3: I then used all of my tricks that I used 260 00:16:57,280 --> 00:17:01,160 Speaker 3: in my research to put that into a bunch of 261 00:17:01,160 --> 00:17:04,199 Speaker 3: different search engines. And I tracked down this guy. His 262 00:17:04,320 --> 00:17:07,879 Speaker 3: name is Peter Madden, and he's a British engineering executive, 263 00:17:08,440 --> 00:17:11,159 Speaker 3: you know, probably similar age. He's a white guy with 264 00:17:11,240 --> 00:17:14,919 Speaker 3: some gray hair, and you know, and and and that 265 00:17:15,119 --> 00:17:21,240 Speaker 3: was presented as you so you know, so it's not 266 00:17:21,600 --> 00:17:27,240 Speaker 3: uncomun for the assistance to make mistakes. And I would 267 00:17:27,320 --> 00:17:31,160 Speaker 3: love to ask Wegmans. You know, if somebody comes up 268 00:17:31,160 --> 00:17:34,680 Speaker 3: to me and escorts me out of the store. Are 269 00:17:34,720 --> 00:17:39,919 Speaker 3: they Are they going to tell me why or do 270 00:17:40,000 --> 00:17:42,560 Speaker 3: I have a chance? You know, are they going to say, 271 00:17:43,000 --> 00:17:46,840 Speaker 3: mister Smith, you know you have to leave? And then 272 00:17:47,119 --> 00:17:48,400 Speaker 3: do I have to produce an ID? 273 00:17:48,800 --> 00:17:48,960 Speaker 4: Right? 274 00:17:49,000 --> 00:17:54,080 Speaker 3: I mean, it creates a lot of uncomfortable situations, and 275 00:17:54,840 --> 00:17:57,280 Speaker 3: you know, there's a lot of civil liberty issues there 276 00:17:57,400 --> 00:18:01,040 Speaker 3: now out of a Wegmans is not the worst thing 277 00:18:01,080 --> 00:18:04,479 Speaker 3: that ever happened to anybody, But it's to me that 278 00:18:04,560 --> 00:18:06,880 Speaker 3: would be that would be a reason that would give 279 00:18:06,960 --> 00:18:10,320 Speaker 3: me pause to go into a store, because I you know, 280 00:18:10,520 --> 00:18:14,960 Speaker 3: I just I'm not sure it's really worth giving that 281 00:18:15,119 --> 00:18:19,080 Speaker 3: up for certain other applications. I do think it's worth 282 00:18:19,200 --> 00:18:22,439 Speaker 3: if the appropriate controls are in place, but I'm not 283 00:18:22,480 --> 00:18:24,399 Speaker 3: sure it's worth it for a grocery store. 284 00:18:25,520 --> 00:18:27,720 Speaker 2: Just a quick question if I could ask, since you 285 00:18:28,320 --> 00:18:32,160 Speaker 2: raised it for me. Of the fifty pictures that came back, 286 00:18:33,640 --> 00:18:35,640 Speaker 2: I've been a public figure for a long time, having 287 00:18:35,680 --> 00:18:38,919 Speaker 2: worked in television in the seventies, eighties and nineties and 288 00:18:38,960 --> 00:18:42,679 Speaker 2: into the two thousands. I also have a Namemake son 289 00:18:42,800 --> 00:18:47,480 Speaker 2: who he looks different than me, but there's some similarities, 290 00:18:47,480 --> 00:18:51,119 Speaker 2: who's also a public figure, having worked in professional baseball. 291 00:18:51,800 --> 00:18:55,000 Speaker 2: Were all the forty nine pictures of me recently? Did 292 00:18:55,000 --> 00:18:56,959 Speaker 2: they go back some years? 293 00:18:57,760 --> 00:19:01,760 Speaker 3: They definitely went back some years. Yes, at least I 294 00:19:01,800 --> 00:19:06,600 Speaker 3: mean looked like probably at least thirty years. Yeah. Yeah, 295 00:19:07,119 --> 00:19:10,560 Speaker 3: definitely a wide span of pictures, a lot of repeats. 296 00:19:10,600 --> 00:19:15,240 Speaker 3: I mean, you've got your professional photos use most often 297 00:19:15,320 --> 00:19:18,679 Speaker 3: those show up multiple times on multiple websites and stuff. 298 00:19:18,680 --> 00:19:21,360 Speaker 3: But yeah, there's quite a big variety. 299 00:19:22,359 --> 00:19:25,639 Speaker 2: Yeah, and none of those pictures appeared on an FBI 300 00:19:25,800 --> 00:19:26,800 Speaker 2: Most Wanted poster. 301 00:19:27,000 --> 00:19:30,639 Speaker 3: Please assure me that I have no comment, no comment. 302 00:19:31,000 --> 00:19:33,879 Speaker 2: Well, you know, I went after FBI agents in my 303 00:19:34,000 --> 00:19:39,320 Speaker 2: television career up here who had unjustly imprisoned people who 304 00:19:39,400 --> 00:19:42,959 Speaker 2: they knew were innocent, and those individuals won more than 305 00:19:42,960 --> 00:19:46,040 Speaker 2: one hundred million dollars in recovery, and they also got 306 00:19:46,080 --> 00:19:49,439 Speaker 2: released from prison as result. So if you told me 307 00:19:49,480 --> 00:19:52,040 Speaker 2: I was on an FBI Most Wanted poster, it wouldn't 308 00:19:52,080 --> 00:19:57,000 Speaker 2: surprise me. That's a true story. If you googled me, 309 00:19:57,119 --> 00:19:57,880 Speaker 2: I know I read. 310 00:19:57,920 --> 00:19:59,080 Speaker 3: I read all about it. 311 00:20:02,359 --> 00:20:05,920 Speaker 2: You did your research. I'm impressed. I'm impressed. Professor Eric 312 00:20:06,000 --> 00:20:10,440 Speaker 2: Learned Miller. We're having some fun. We have callers up already, 313 00:20:10,520 --> 00:20:12,719 Speaker 2: and we're going to get to them, I promise, and 314 00:20:12,760 --> 00:20:15,359 Speaker 2: we have some room for you. One at six, one seven, 315 00:20:15,400 --> 00:20:18,560 Speaker 2: two thirty and one at six one, seven, nine three, 316 00:20:18,640 --> 00:20:20,960 Speaker 2: one tenth third. We're gonna get to him, Rick and 317 00:20:21,040 --> 00:20:24,280 Speaker 2: Bob and Moore right here on Nightside. Let's have some 318 00:20:24,400 --> 00:20:26,320 Speaker 2: fun with this, because I gotta tell you there is 319 00:20:26,359 --> 00:20:28,879 Speaker 2: a little bit of an Elvis Huxley brave New World 320 00:20:29,000 --> 00:20:32,080 Speaker 2: sense of this, all of this stuff that we read 321 00:20:32,080 --> 00:20:34,800 Speaker 2: about fifty years ago and said, nah, that'll never happen. 322 00:20:35,119 --> 00:20:37,320 Speaker 2: A lot of the stuff we read about didn't happen, 323 00:20:37,359 --> 00:20:41,960 Speaker 2: flying cars and stuff like that. But this facial recognition, uh, 324 00:20:42,000 --> 00:20:44,080 Speaker 2: it can. It can be used for the good are 325 00:20:44,280 --> 00:20:48,080 Speaker 2: the not so good. And I'm delighted that Professor Learned 326 00:20:48,119 --> 00:20:50,879 Speaker 2: Miller uh is spending the time with us because he 327 00:20:50,960 --> 00:20:53,720 Speaker 2: has been in on the ground floor. He knows the 328 00:20:53,840 --> 00:20:56,200 Speaker 2: positives and he knows the negatives. And we'll be back 329 00:20:56,240 --> 00:20:58,600 Speaker 2: on Nightside right after this news at the bottom of 330 00:20:58,640 --> 00:21:01,160 Speaker 2: the hour, You're. 331 00:21:01,000 --> 00:21:05,640 Speaker 1: On Night Side with Dan Ray on Boston's news Radio. 332 00:21:06,600 --> 00:21:11,520 Speaker 2: We're talking with Professor Eric Learnid Miller. He's a professor, 333 00:21:11,560 --> 00:21:16,680 Speaker 2: he's actually the chair of the faculty. Professor and chair 334 00:21:16,720 --> 00:21:19,720 Speaker 2: of the faculty at the UMass Amherst Manning College of 335 00:21:19,760 --> 00:21:23,720 Speaker 2: Information and Computer Sciences. By the way, I did find 336 00:21:23,720 --> 00:21:27,600 Speaker 2: a picture of this guy, Peter Madden, my doppeler Dopplinger. 337 00:21:28,400 --> 00:21:31,439 Speaker 2: He's a governor and a member of a senior leadership 338 00:21:31,520 --> 00:21:35,600 Speaker 2: team at at something called the Reeds School. Is that 339 00:21:35,640 --> 00:21:36,320 Speaker 2: the guy you found? 340 00:21:36,480 --> 00:21:40,000 Speaker 3: Because this is the same guy. Yeah, she's that's a 341 00:21:40,080 --> 00:21:44,240 Speaker 3: new position for him. And he's also been like a 342 00:21:44,320 --> 00:21:48,119 Speaker 3: construction engineering executive and done a few other things. 343 00:21:48,119 --> 00:21:51,320 Speaker 2: But he looks like he's a handsome devil. Looks like 344 00:21:51,359 --> 00:21:52,040 Speaker 2: a movie star. 345 00:21:54,200 --> 00:21:59,680 Speaker 3: Yeah, I know you. You can take that as a compliment. 346 00:22:00,880 --> 00:22:05,000 Speaker 2: Very distinguished looking gentlemen. It'd be embarrassed if he ever 347 00:22:05,040 --> 00:22:08,880 Speaker 2: found out that somehow he was linked to me through 348 00:22:08,880 --> 00:22:13,280 Speaker 2: facial But I could see the resemblance. Maybe I don't know, Yeah. 349 00:22:13,400 --> 00:22:16,680 Speaker 3: I mean to me, it was you know, it wasn't 350 00:22:16,720 --> 00:22:19,439 Speaker 3: in the top ten pictures, but it was down the 351 00:22:19,520 --> 00:22:22,000 Speaker 3: list and it was presented as the picture of you. 352 00:22:22,280 --> 00:22:24,520 Speaker 3: So there you go. 353 00:22:25,440 --> 00:22:27,800 Speaker 2: Very cool, Very cool. Let's go to phone calls. Okay, 354 00:22:27,840 --> 00:22:31,000 Speaker 2: We're going to start it off with Rick in Rockland, 355 00:22:31,080 --> 00:22:34,760 Speaker 2: Massachusetts lit. Closer to Home, Rick, Welcome, you're on Professor 356 00:22:34,880 --> 00:22:40,359 Speaker 2: Eric Leernard Miller of UMass Amherst, and we're talking about 357 00:22:40,440 --> 00:22:44,120 Speaker 2: the fact that Wegman's now in some stores, not all, 358 00:22:44,520 --> 00:22:47,720 Speaker 2: its starting to use facial recognition technology. What do you think, Rick, 359 00:22:48,960 --> 00:22:49,280 Speaker 2: oh be. 360 00:22:49,320 --> 00:22:52,040 Speaker 6: Even guys, I think it's very interesting. It's the second 361 00:22:52,400 --> 00:22:55,480 Speaker 6: small retail store that I've heard using nothing of technology. 362 00:22:56,040 --> 00:22:59,040 Speaker 6: I know that Regal marketplace equipment is using it, and 363 00:22:59,040 --> 00:23:01,960 Speaker 6: they're really small. I can't figure out why they'd want 364 00:23:01,960 --> 00:23:04,080 Speaker 6: to do that. Like you guys are talking about, maybe 365 00:23:04,080 --> 00:23:06,440 Speaker 6: if someone was pressed past, maybe if someone needs to 366 00:23:06,520 --> 00:23:11,119 Speaker 6: use the bathroom, or maybe it's to prevent people use 367 00:23:11,200 --> 00:23:13,080 Speaker 6: the bathroom for other things those lose the bathroom, you 368 00:23:13,119 --> 00:23:15,800 Speaker 6: know what I mean, like people that have substance. 369 00:23:15,400 --> 00:23:19,320 Speaker 2: To use problems. Yeah, I get it, But you would 370 00:23:19,400 --> 00:23:23,760 Speaker 2: think that even if they just had people at a supermarket, 371 00:23:23,800 --> 00:23:26,040 Speaker 2: they had people on the floor. And again my expert 372 00:23:26,040 --> 00:23:30,879 Speaker 2: here is Professor Learnon Miller, So my my quick comment is, 373 00:23:30,920 --> 00:23:35,040 Speaker 2: you would think they could accomplish whatever traffic control in 374 00:23:35,119 --> 00:23:38,400 Speaker 2: and out of the bathroom they wanted for non patrons, 375 00:23:38,800 --> 00:23:41,159 Speaker 2: and particularly people who come in and look like they're 376 00:23:41,960 --> 00:23:44,320 Speaker 2: you know, they're they're in they're in some some form 377 00:23:44,359 --> 00:23:50,280 Speaker 2: of distress, Doctor Lernon Miller, Professor Lernon Miller, your comment 378 00:23:50,359 --> 00:23:53,120 Speaker 2: to say had to Rick from Rockland and your your 379 00:23:53,320 --> 00:23:55,600 Speaker 2: observations on his observation. 380 00:23:56,640 --> 00:24:01,920 Speaker 3: Yeah, it's it's yeah. I I certainly don't know why 381 00:24:01,960 --> 00:24:05,840 Speaker 3: they're doing it, and I share all the same questions 382 00:24:05,840 --> 00:24:09,880 Speaker 3: that you're raising. I mean, you know, maybe one thing 383 00:24:10,000 --> 00:24:14,320 Speaker 3: that is slightly relevant is that in some cases, I 384 00:24:14,359 --> 00:24:17,399 Speaker 3: think it was maybe Madison Square Garden that was starting 385 00:24:17,440 --> 00:24:21,640 Speaker 3: to use but you know, forgive me if I got 386 00:24:21,680 --> 00:24:24,520 Speaker 3: the wrong location, but I think there was a large 387 00:24:24,600 --> 00:24:29,520 Speaker 3: venue that was using face recognition software and some of 388 00:24:29,560 --> 00:24:36,240 Speaker 3: the management in one of these locations started identifying people 389 00:24:37,119 --> 00:24:41,240 Speaker 3: that had been involved in a legal case against them 390 00:24:41,320 --> 00:24:48,000 Speaker 3: and preventing them from coming in to coming into the venue. 391 00:24:48,560 --> 00:24:52,879 Speaker 3: And so, you know, there's so many questions that you 392 00:24:52,920 --> 00:24:57,280 Speaker 3: know arise there when you're claiming it's for one use, 393 00:24:57,359 --> 00:25:01,760 Speaker 3: but then people get tempted to use it for other things. 394 00:25:01,840 --> 00:25:04,560 Speaker 3: And I think, you know, there are ways to try 395 00:25:04,600 --> 00:25:08,879 Speaker 3: to prevent this by having really clear rules and having 396 00:25:08,960 --> 00:25:14,320 Speaker 3: clear consequences for the misuse of those rules. But unfortunately, 397 00:25:15,680 --> 00:25:19,400 Speaker 3: even if there are rules, they're usually unenforceable. And there's 398 00:25:19,440 --> 00:25:24,560 Speaker 3: usually no consequences for them, and so there have been 399 00:25:24,680 --> 00:25:28,359 Speaker 3: quite a few cases. Let me give you one little 400 00:25:28,400 --> 00:25:31,320 Speaker 3: story that I think is relevant to this whole discussion. 401 00:25:31,720 --> 00:25:35,520 Speaker 3: So there was a gentleman in Detroit who we were 402 00:25:35,520 --> 00:25:39,760 Speaker 3: talking about jewelry stores earlier, and he was arrested for 403 00:25:39,920 --> 00:25:45,600 Speaker 3: robbing a jewelry store based on a face id between 404 00:25:45,800 --> 00:25:48,480 Speaker 3: the person who robbed the store that would turned out 405 00:25:48,520 --> 00:25:52,720 Speaker 3: to not be him and his own driver's license picture. 406 00:25:53,600 --> 00:25:57,720 Speaker 3: And the Detroit the police department had a rule that 407 00:25:57,880 --> 00:26:02,679 Speaker 3: said a face recognition alone can never be used for 408 00:26:02,760 --> 00:26:07,399 Speaker 3: an arrest. You have to have corroborating evidence. But in 409 00:26:07,440 --> 00:26:11,600 Speaker 3: that case, they ignored the rule and they arrested him anyway, 410 00:26:12,000 --> 00:26:14,199 Speaker 3: and he was put in jail for the weekend. He 411 00:26:14,280 --> 00:26:17,119 Speaker 3: was arrested in front of his two little girls, He 412 00:26:17,280 --> 00:26:20,840 Speaker 3: missed a day at work, his employer thought he'd been arrested. 413 00:26:20,960 --> 00:26:24,960 Speaker 3: You know, all kinds of stuff because even though they 414 00:26:25,000 --> 00:26:27,800 Speaker 3: had these rules or I was supposed to be used, 415 00:26:28,280 --> 00:26:32,080 Speaker 3: they didn't follow the rules. And you know, one of 416 00:26:32,119 --> 00:26:34,640 Speaker 3: the one of the things that's often lacking from these 417 00:26:34,680 --> 00:26:39,800 Speaker 3: things is consequences if you don't follow the rules. You know, 418 00:26:39,960 --> 00:26:43,040 Speaker 3: for example, if you're I talk a lot about drug 419 00:26:43,080 --> 00:26:47,680 Speaker 3: regulations because they're much better developed. Like, if you're a pharmacist, 420 00:26:48,240 --> 00:26:53,480 Speaker 3: you break the rules for distributing drugs and your pharmacy, 421 00:26:53,840 --> 00:26:57,639 Speaker 3: you can lose your license. But there's no such rules 422 00:26:57,680 --> 00:27:01,879 Speaker 3: like that with face recognition, which means there's very little 423 00:27:01,960 --> 00:27:06,000 Speaker 3: incentive for people to really stick to the rules when 424 00:27:06,000 --> 00:27:08,840 Speaker 3: they don't feel like it. So you know, that's one 425 00:27:08,880 --> 00:27:12,720 Speaker 3: of the problems. And whatever they're using it for, I 426 00:27:13,040 --> 00:27:16,440 Speaker 3: don't know, but I would have those concerns. I guess 427 00:27:16,800 --> 00:27:17,520 Speaker 3: real quick. 428 00:27:17,359 --> 00:27:19,640 Speaker 2: Question for me, and that is, if you found out 429 00:27:19,720 --> 00:27:22,879 Speaker 2: that there was a supermarket or a store that you 430 00:27:23,000 --> 00:27:28,080 Speaker 2: normally would frequent using facial recognition, would that change your 431 00:27:28,720 --> 00:27:30,600 Speaker 2: decision to go in and out of that store on 432 00:27:30,640 --> 00:27:31,960 Speaker 2: a regular basis? 433 00:27:32,600 --> 00:27:35,560 Speaker 6: Yeah, I would probably try to avoid it if possible. 434 00:27:35,640 --> 00:27:37,639 Speaker 6: For you know, I bet a question this might be 435 00:27:37,720 --> 00:27:43,439 Speaker 6: better very How can I minimize my risk or produce 436 00:27:43,480 --> 00:27:45,600 Speaker 6: their possibility of just looking at my face? 437 00:27:45,840 --> 00:27:47,680 Speaker 4: Is there certain. 438 00:27:47,359 --> 00:27:50,000 Speaker 6: Types of patterns I could wear or anything I can 439 00:27:50,080 --> 00:27:50,639 Speaker 6: kind of trick it. 440 00:27:50,640 --> 00:27:51,080 Speaker 4: A little bit. 441 00:27:52,040 --> 00:27:53,399 Speaker 2: I guess you could wear a face mask. 442 00:27:53,600 --> 00:27:56,240 Speaker 3: You could certainly if you put on a baseball cap, 443 00:27:56,359 --> 00:27:58,680 Speaker 3: you're definitely going to cut off a lot of the angles, 444 00:27:58,720 --> 00:28:02,679 Speaker 3: you know, and there's you know, people always have examples 445 00:28:02,720 --> 00:28:06,359 Speaker 3: online of like how face recognition can work even with 446 00:28:06,440 --> 00:28:10,679 Speaker 3: disguises and all that, but usually it doesn't work anywhere 447 00:28:10,760 --> 00:28:13,640 Speaker 3: near as well. And so just wearing a baseball cap 448 00:28:13,680 --> 00:28:17,040 Speaker 3: and a pair of sunglasses is going to dramatically reduce 449 00:28:17,200 --> 00:28:18,400 Speaker 3: its capabilities. 450 00:28:19,840 --> 00:28:20,080 Speaker 5: You know. 451 00:28:20,240 --> 00:28:24,120 Speaker 3: Of course you might be mismatched with somebody by doing that, 452 00:28:24,320 --> 00:28:27,879 Speaker 3: and so it's not clear that's actually going to help you. 453 00:28:28,320 --> 00:28:32,280 Speaker 3: I mean, I would be really interested to hear from Wegmans, 454 00:28:32,320 --> 00:28:37,280 Speaker 3: like what do they do if they claim that you 455 00:28:37,560 --> 00:28:42,560 Speaker 3: match somebody who previously stole something? Do they scort you out? 456 00:28:42,680 --> 00:28:45,440 Speaker 3: Do they ask you for an ID? I mean, it's 457 00:28:45,520 --> 00:28:49,120 Speaker 3: so awkward and so off putting if they make a mistake, 458 00:28:49,840 --> 00:28:52,720 Speaker 3: I'm kind of amazed it would be worth it from 459 00:28:52,720 --> 00:28:55,880 Speaker 3: a customer relations point of view, but maybe it is. 460 00:28:57,240 --> 00:29:00,680 Speaker 2: Rick appreciate you call very much, good, good questions comments. 461 00:29:00,720 --> 00:29:02,520 Speaker 2: I got a break here and I got a couple 462 00:29:02,640 --> 00:29:05,640 Speaker 2: more calls coming in. Thanks very much, first time calling 463 00:29:05,680 --> 00:29:06,680 Speaker 2: my show record now. 464 00:29:08,520 --> 00:29:09,840 Speaker 3: Before well keep calling. 465 00:29:10,040 --> 00:29:12,560 Speaker 2: Really appreciate. Happy new you to you and yours. Thank 466 00:29:12,560 --> 00:29:15,200 Speaker 2: you so much. I have a great and I would 467 00:29:15,200 --> 00:29:17,960 Speaker 2: take a quick break coming back at Bob and California, 468 00:29:18,160 --> 00:29:21,120 Speaker 2: lesson New Hampshire. As I told you, Professor, we do 469 00:29:21,200 --> 00:29:23,520 Speaker 2: get around, that's for sure, and this has been a 470 00:29:23,560 --> 00:29:27,480 Speaker 2: fun topic. I appreciate the time you've taken with us tonight, 471 00:29:27,520 --> 00:29:29,520 Speaker 2: and I promise i'll let you go. By ten o'clock. 472 00:29:29,600 --> 00:29:32,320 Speaker 2: We'll be back on night Side with my guest, Professor 473 00:29:32,480 --> 00:29:35,560 Speaker 2: Eric Learned Miller of University of Massachusetts. He is an 474 00:29:35,600 --> 00:29:39,040 Speaker 2: expert in facial recognition. I knew that, but he certainly 475 00:29:39,080 --> 00:29:42,480 Speaker 2: has proven it this evening, and it's been a great guest, 476 00:29:42,520 --> 00:29:44,440 Speaker 2: and I really appreciate his time. We'll be back on 477 00:29:44,520 --> 00:29:48,600 Speaker 2: Nightside with more calls for Professor Eric Learned Miller of 478 00:29:48,640 --> 00:29:49,320 Speaker 2: you mass. 479 00:29:49,080 --> 00:29:54,280 Speaker 1: Amherst night Side with Dan Ray. I'm delling you bes 480 00:29:54,800 --> 00:29:56,240 Speaker 1: Boston's News Radio. 481 00:29:56,760 --> 00:29:59,920 Speaker 2: We're talking about grocery stores Wegmans in some of the 482 00:30:00,160 --> 00:30:05,160 Speaker 2: stores now using facial recognition. We have read the statement. 483 00:30:05,240 --> 00:30:07,840 Speaker 2: We also have as a very special guest, Professor Eric 484 00:30:07,920 --> 00:30:10,640 Speaker 2: Lorded Miller of you Masks. Let me go to Bob 485 00:30:10,680 --> 00:30:13,640 Speaker 2: in California. Hey Bob, welcome back, Happy New Year. You 486 00:30:13,680 --> 00:30:16,280 Speaker 2: were on with Professor Eric Loarded Miller of you mask 487 00:30:16,320 --> 00:30:16,640 Speaker 2: go right. 488 00:30:16,560 --> 00:30:20,440 Speaker 5: Ahead, Bob, Yeah, Dan, thank you, and hello, professor, Hello 489 00:30:20,480 --> 00:30:25,160 Speaker 5: to both of you. Yes, I really respectfully say, I 490 00:30:25,200 --> 00:30:29,200 Speaker 5: don't think either one of you really understands or grass 491 00:30:30,040 --> 00:30:34,560 Speaker 5: the reality that even grocery stores are facing in the 492 00:30:34,640 --> 00:30:38,479 Speaker 5: real life in terms of business. Okay. I was in 493 00:30:38,560 --> 00:30:42,560 Speaker 5: actually two weeks ago in Nevada, in Henderson nice area, 494 00:30:43,080 --> 00:30:45,480 Speaker 5: one of the Kroger I won't name which branch or, 495 00:30:45,520 --> 00:30:49,440 Speaker 5: one of the Kroger grocery stores, and they're locking up, 496 00:30:49,480 --> 00:30:51,760 Speaker 5: as a lot of stores are now, a lot of 497 00:30:51,800 --> 00:30:54,240 Speaker 5: items are locked up. Sure, I was trying to get 498 00:30:54,280 --> 00:30:57,480 Speaker 5: a forty dollars bottle of alcohol. I had to hit 499 00:30:57,520 --> 00:31:01,400 Speaker 5: the button. Manager came over talking for a couple of minutes, 500 00:31:02,040 --> 00:31:04,600 Speaker 5: and I'm like, is this really a problem, And he 501 00:31:04,640 --> 00:31:08,800 Speaker 5: shook his head and he goes, try thousands of dollars, 502 00:31:09,120 --> 00:31:12,080 Speaker 5: he goes thousands of dollars every week, he goes the 503 00:31:12,160 --> 00:31:15,600 Speaker 5: next per store, he goes. It's millions and millions for 504 00:31:15,680 --> 00:31:20,040 Speaker 5: our company every year. And he reminded me, and again, 505 00:31:20,080 --> 00:31:21,520 Speaker 5: I don't know if either one of you know this. 506 00:31:21,640 --> 00:31:24,720 Speaker 5: I don't think so grocery stores operate on a two 507 00:31:24,840 --> 00:31:25,840 Speaker 5: percent profit. 508 00:31:26,080 --> 00:31:29,320 Speaker 2: That's I knew that that that I do they have. 509 00:31:29,800 --> 00:31:33,240 Speaker 2: They have a huge volume of a huge volume. But 510 00:31:33,280 --> 00:31:34,880 Speaker 2: that is that is the profit margin. 511 00:31:35,800 --> 00:31:38,760 Speaker 5: That's the right. So the bottom line is if you're talking, 512 00:31:39,240 --> 00:31:42,000 Speaker 5: let's just call it a number. Let's call it five 513 00:31:42,040 --> 00:31:45,560 Speaker 5: thousand dollars. This is a number per week. Ninety eight 514 00:31:45,680 --> 00:31:49,320 Speaker 5: percent of that the store has to eat that cost. 515 00:31:49,720 --> 00:31:52,240 Speaker 5: They have to eat that and guess what they pass 516 00:31:52,320 --> 00:31:56,040 Speaker 5: it on to the customers. The customers are paying more prices, 517 00:31:56,160 --> 00:31:59,560 Speaker 5: no question they have. They have an absolute right any 518 00:31:59,600 --> 00:32:04,440 Speaker 5: private is, including grocery stores. It's their business, okay. And 519 00:32:04,520 --> 00:32:07,080 Speaker 5: as long as they put signs out in front of 520 00:32:07,080 --> 00:32:11,440 Speaker 5: the store, video surveillance is underway. Here are your subject 521 00:32:11,440 --> 00:32:15,200 Speaker 5: of video surveillance. That's it, period, end of the story. 522 00:32:15,520 --> 00:32:18,080 Speaker 5: People go in knowing that or don't coin Yeah. 523 00:32:18,720 --> 00:32:21,760 Speaker 2: Well again, I'm with you, Bob. I think you're correct. 524 00:32:21,840 --> 00:32:24,120 Speaker 2: That's what Wegmans has done here, professor. 525 00:32:25,520 --> 00:32:32,840 Speaker 3: Feel First of all, I appreciate, uh, you know, the 526 00:32:32,960 --> 00:32:35,840 Speaker 3: extra information. I mean, I have a lot to learn 527 00:32:36,000 --> 00:32:41,280 Speaker 3: about retail, no question about it. And and it's I 528 00:32:41,560 --> 00:32:43,880 Speaker 3: you know, as I was saying a little bit earlier, 529 00:32:44,600 --> 00:32:47,360 Speaker 3: you know, they are trying to save money. They're trying 530 00:32:47,360 --> 00:32:51,680 Speaker 3: to reduce staff. I'm sure, and and that totally makes 531 00:32:51,720 --> 00:32:54,560 Speaker 3: sense to me. I think, you know one interesting question, 532 00:32:54,720 --> 00:32:58,760 Speaker 3: like you talked about video surrounds, which could people be 533 00:32:58,920 --> 00:33:03,640 Speaker 3: people just recording video or watching it, you know, in 534 00:33:03,760 --> 00:33:09,240 Speaker 3: like a control room or something. And I think there 535 00:33:09,280 --> 00:33:13,920 Speaker 3: aren't too many people at least I'm certainly not somebody 536 00:33:13,920 --> 00:33:18,600 Speaker 3: who objects to a store having video surveillance to try 537 00:33:18,640 --> 00:33:22,479 Speaker 3: to reduce theft. I mean that's widespread. It's in the 538 00:33:22,480 --> 00:33:25,160 Speaker 3: convenience store I go to down the street. You know, 539 00:33:25,880 --> 00:33:29,320 Speaker 3: that's everywhere. And I think people have accepted that. But 540 00:33:30,480 --> 00:33:36,520 Speaker 3: you know, this is more about like this adds another layer, right, 541 00:33:36,920 --> 00:33:42,040 Speaker 3: which is like, okay, there was there was a person 542 00:33:42,360 --> 00:33:45,480 Speaker 3: in the paper who was arrested for something, and our 543 00:33:45,560 --> 00:33:50,120 Speaker 3: customer appears to match that same person, right, even though 544 00:33:50,120 --> 00:33:53,880 Speaker 3: we haven't seen them do anything bad, they appear to 545 00:33:54,000 --> 00:33:58,600 Speaker 3: match that person. And the problem is that the technology 546 00:33:58,720 --> 00:34:03,880 Speaker 3: sometimes makes mistake So unlike unlike a surveillance camera where 547 00:34:03,880 --> 00:34:07,480 Speaker 3: you watch somebody put a fit of bourbon in their 548 00:34:07,560 --> 00:34:12,200 Speaker 3: coat and you go grasp them, right, then this is 549 00:34:12,400 --> 00:34:18,799 Speaker 3: about drawing connections with a semi flawed system. So it's 550 00:34:18,840 --> 00:34:23,080 Speaker 3: another layer, and you know, I feel for the owners. 551 00:34:23,120 --> 00:34:26,120 Speaker 3: They're trying to reduce that bed totally makes sense. 552 00:34:26,200 --> 00:34:28,440 Speaker 2: Absolutely, Hey Bob, thanks, Okay, today. 553 00:34:28,760 --> 00:34:33,760 Speaker 5: John, I wanted to make another point Dan if I did. Okay, professor, Okay, 554 00:34:34,080 --> 00:34:34,720 Speaker 5: if you could. 555 00:34:34,560 --> 00:34:36,239 Speaker 2: Be real quick, because I got one of the call 556 00:34:36,320 --> 00:34:37,240 Speaker 2: I want to give them add. 557 00:34:37,440 --> 00:34:40,800 Speaker 5: Again what what the professor said? I get that, professor, 558 00:34:40,840 --> 00:34:43,680 Speaker 5: but that's exceptions and the person would heavily go right. 559 00:34:43,719 --> 00:34:45,960 Speaker 5: So the bottom line is if somebody goes in the 560 00:34:46,040 --> 00:34:49,480 Speaker 5: private property and they know that they're being video surveillance, 561 00:34:49,560 --> 00:34:52,120 Speaker 5: they know that that business has the right to do 562 00:34:52,160 --> 00:34:55,120 Speaker 5: whatever they want to do with that video. It's there, right, 563 00:34:55,160 --> 00:35:03,920 Speaker 5: you're on private property that yeah, yeah, yeah, So you 564 00:35:03,960 --> 00:35:07,720 Speaker 5: know this this whole thing about you know, the mistakes 565 00:35:07,760 --> 00:35:08,360 Speaker 5: happen once. 566 00:35:08,200 --> 00:35:08,600 Speaker 3: In a while. 567 00:35:08,800 --> 00:35:11,840 Speaker 5: People can have legal recourse to that. Again, you're on 568 00:35:11,920 --> 00:35:14,839 Speaker 5: private property, Okay, got definitely trying to protect it. 569 00:35:14,920 --> 00:35:16,960 Speaker 2: But Bob, I think you made that point. You made 570 00:35:17,000 --> 00:35:18,279 Speaker 2: it well, But I got to run. I got to 571 00:35:18,280 --> 00:35:20,720 Speaker 2: give one other minute to one other caller who's waited 572 00:35:20,840 --> 00:35:22,680 Speaker 2: just as long. Thank you much. Let me go to 573 00:35:22,760 --> 00:35:25,319 Speaker 2: Less in New Hampshire. Less. I hate to tell you 574 00:35:25,320 --> 00:35:27,080 Speaker 2: you got less than a minute. What would you like 575 00:35:27,120 --> 00:35:28,160 Speaker 2: to say, go right ahead. 576 00:35:29,400 --> 00:35:31,840 Speaker 4: That's okay. Thanks for taking my call. Dan, Thank you 577 00:35:31,880 --> 00:35:33,800 Speaker 4: a long time listener, first time caller. 578 00:35:34,800 --> 00:35:37,560 Speaker 2: To quick round of applause from our studio audience here 579 00:35:37,320 --> 00:35:40,360 Speaker 2: a virtual studio audience. Go right ahead, go ahead. 580 00:35:40,719 --> 00:35:42,279 Speaker 4: You know, one of the things that I haven't heard 581 00:35:42,360 --> 00:35:46,200 Speaker 4: yet is the idea that the facial recognition could be 582 00:35:46,320 --> 00:35:51,279 Speaker 4: used to help identify criminals that are are are part 583 00:35:51,280 --> 00:35:53,719 Speaker 4: of scams. You know a lot of people have been 584 00:35:53,880 --> 00:35:57,320 Speaker 4: burned by gift card scams and things like that, online scams, 585 00:35:57,760 --> 00:36:01,880 Speaker 4: and you know, the defenders that go into stores with 586 00:36:02,320 --> 00:36:06,840 Speaker 4: stolen credit cards, ebt cards for grocery stores. You know, 587 00:36:06,920 --> 00:36:09,000 Speaker 4: those things happen all the time, but there's really no 588 00:36:09,160 --> 00:36:14,440 Speaker 4: way of knowing necessarily who has an EBT card that's 589 00:36:14,680 --> 00:36:18,040 Speaker 4: been stolen. So that's the facial recognition is probably, in 590 00:36:18,080 --> 00:36:21,960 Speaker 4: my opinion, one way of being able to put a 591 00:36:22,000 --> 00:36:25,120 Speaker 4: face to something that's that that shouldn't be happening. So 592 00:36:25,480 --> 00:36:28,160 Speaker 4: I don't know if that I think facial recognition is 593 00:36:28,200 --> 00:36:31,400 Speaker 4: here to stay. It's the next you know, it's like 594 00:36:31,480 --> 00:36:34,520 Speaker 4: talking about aian. These things are just going to be 595 00:36:34,560 --> 00:36:37,200 Speaker 4: a part of our life and there's almost nothing they 596 00:36:37,200 --> 00:36:38,000 Speaker 4: can do about. 597 00:36:37,800 --> 00:36:41,040 Speaker 2: It unless you make great points. I think it's the 598 00:36:41,080 --> 00:36:43,960 Speaker 2: next frontier. I'm flat out of time, boy. I wish 599 00:36:44,040 --> 00:36:46,200 Speaker 2: you called earlier next time. I want you to call 600 00:36:46,280 --> 00:36:48,799 Speaker 2: early because you sound like a great caller. Thank you 601 00:36:48,840 --> 00:36:50,239 Speaker 2: so much. We'll talk to you soon. 602 00:36:50,440 --> 00:36:52,120 Speaker 4: Thank you, good and welcome. 603 00:36:52,200 --> 00:36:56,960 Speaker 2: Thank you, professor, Professor Eric Learned Miller of you, mass Amherst. 604 00:36:56,960 --> 00:36:59,200 Speaker 2: This was a great hour. Thank you for your time, 605 00:36:59,360 --> 00:37:01,840 Speaker 2: Thank you for your expertise. I couldn't have done it 606 00:37:01,880 --> 00:37:04,000 Speaker 2: without you, sir, and I really do appreciate it. Hope 607 00:37:04,040 --> 00:37:04,800 Speaker 2: to meet you someday. 608 00:37:04,960 --> 00:37:07,080 Speaker 3: Yeah. And you know, I just want to say I 609 00:37:07,080 --> 00:37:11,080 Speaker 3: think all the college had good points. I you know, 610 00:37:11,440 --> 00:37:16,080 Speaker 3: I appreciate the need of these, you know, of store 611 00:37:16,080 --> 00:37:18,560 Speaker 3: owner to try to you know, do what they can 612 00:37:18,760 --> 00:37:22,680 Speaker 3: to to nip the stuff in the butd and you know, 613 00:37:22,840 --> 00:37:25,920 Speaker 3: we just want to do it thoughtfully and you know, 614 00:37:26,200 --> 00:37:30,279 Speaker 3: keep people's rights in mind. And it probably is here 615 00:37:30,320 --> 00:37:31,920 Speaker 3: to stay, no no doubt. 616 00:37:32,080 --> 00:37:35,080 Speaker 2: Professor. Thanks very much. I'd like to retain the option 617 00:37:35,200 --> 00:37:37,000 Speaker 2: to get back to it at some point in this 618 00:37:37,040 --> 00:37:40,319 Speaker 2: because this was a really interesting hour. And it'll be 619 00:37:40,360 --> 00:37:44,600 Speaker 2: posted at our website on Nightside and Demand and you 620 00:37:44,640 --> 00:37:46,799 Speaker 2: get a chance to re listen to it tomorrow or 621 00:37:46,960 --> 00:37:49,640 Speaker 2: circulate it amongst anyone you like at U mass amhers 622 00:37:49,640 --> 00:37:52,960 Speaker 2: because I think you represented the school tremendously and I 623 00:37:53,000 --> 00:37:55,600 Speaker 2: hope President Mehan has a chance to hear it. Okay, 624 00:37:55,680 --> 00:37:56,479 Speaker 2: thank you so much. 625 00:37:56,680 --> 00:37:58,480 Speaker 3: All right, well, thanks for having me as a guest. 626 00:37:58,840 --> 00:38:02,000 Speaker 2: All right, thank you for Professor Eric lorded Miller. Of 627 00:38:02,040 --> 00:38:04,600 Speaker 2: you mass Amherst. We're tight to the break. Here comes 628 00:38:04,640 --> 00:38:06,600 Speaker 2: the newscast. Sorry Robert in a little late, but I 629 00:38:06,640 --> 00:38:07,520 Speaker 2: want to get everyone in