WEBVTT 00:00.329 --> 00:03.246 (electronic music) 00:10.299 --> 00:12.666 ^- We're moving more and more towards autonomy, 00:12.666 --> 00:15.219 ^both within the Air Force and and within society. 00:15.219 --> 00:17.401 ^In those discussions, you'll see most people 00:17.401 --> 00:19.645 talk about this process of trust. 00:19.645 --> 00:21.793 This issue of trust is foundational 00:21.793 --> 00:23.724 to our acceptance of autonomy. 00:23.724 --> 00:25.196 The idea of human-machine trust 00:25.196 --> 00:27.627 ^is the notion of how we develop reliance 00:27.627 --> 00:29.812 ^and tensions with machines. 00:29.812 --> 00:31.513 So we do this with people all the time. 00:31.513 --> 00:34.032 We study this process with machines. 00:34.032 --> 00:38.025 ^- We seek to help systems develop calibrated trust. 00:38.025 --> 00:40.250 ^So that the human operators 00:40.250 --> 00:43.256 ^have appropriate levels of trust in systems 00:43.256 --> 00:46.224 ^without under trusting or over trusting it. 00:46.224 --> 00:49.180 ^- So when we're evaluating these reliants and tensions, 00:49.180 --> 00:51.848 there's a number of different factors that come into play. 00:51.848 --> 00:54.645 You could consider your dispositions or your personality. 00:54.645 --> 00:57.071 Things like your preferences and experiences. 00:57.071 --> 00:58.580 Your biases that you actually bring 00:58.580 --> 01:00.373 to a situation with you. 01:00.373 --> 01:02.631 All the way to more machine characteristics, 01:02.631 --> 01:04.548 things like reliability, 01:04.548 --> 01:07.980 how familiar is the interface or the technology. 01:07.980 --> 01:10.244 - I love the people and the kind of work that we're doing. 01:10.244 --> 01:11.703 I love that we're doing applied work 01:11.703 --> 01:14.093 that really can be meaningful, 01:14.093 --> 01:15.228 that can really have an impact 01:15.228 --> 01:18.061 on war fighters' lives and safety. 01:18.373 --> 01:20.301 ^It's really interesting work, 01:20.301 --> 01:23.101 ^we're working on a lot of different projects, 01:23.101 --> 01:25.413 and the work really can make a difference. 01:25.413 --> 01:28.733 In some cases we are literally saving lives. 01:28.733 --> 01:29.900 - Hi everyone. 01:30.307 --> 01:33.224 My name is Baxter and I am a robot. 01:33.735 --> 01:36.205 - I'd say my favorite part of working for AFRL 01:36.205 --> 01:38.439 is the fact that we operate at the chasm 01:38.439 --> 01:41.808 of where basic science meets applied need. 01:41.808 --> 01:44.074 With the DOD in the research area, 01:44.074 --> 01:47.324 we're never going to get better applied problems to work on. 01:47.324 --> 01:49.555 Our military has a great deal of needs 01:49.555 --> 01:51.559 and it's an exceptional military, 01:51.559 --> 01:53.699 so being able to improve operations 01:53.699 --> 01:55.227 ^for the military in any way 01:55.227 --> 01:57.993 ^is really value added from my perspective. 01:57.993 --> 01:59.346 It's also really exciting 01:59.346 --> 02:01.676 anytime we can get out and talk to war fighters. 02:01.676 --> 02:04.000 We get to get out there and hear from the pilots 02:04.000 --> 02:05.536 about what some of their concerns are 02:05.536 --> 02:06.880 with the systems that they're using, 02:06.880 --> 02:09.310 feed that information back into the system 02:09.310 --> 02:11.722 so that possible improvements can be made. 02:11.722 --> 02:13.205 - If they under trust a system, 02:13.205 --> 02:15.612 they're not going to use it when it may be effective. 02:15.612 --> 02:16.541 If they over trust it, 02:16.541 --> 02:18.626 they may over rely upon it, 02:18.626 --> 02:20.040 particularly in situations where 02:20.040 --> 02:22.106 maybe they shouldn't trust it. 02:22.106 --> 02:23.459 - It's never about blind trust. 02:23.459 --> 02:25.343 It's never about generating more trust 02:25.343 --> 02:26.706 when it's not warranted. 02:26.706 --> 02:29.838 It's about generating appropriate reliance on technology. 02:29.838 --> 02:32.253 And that's the core of what we do here. 02:32.253 --> 02:35.170 (electronic music)