Dava Newman: Role of AI in Space Travel
00:00:00.000 | - I love artificial intelligence,
00:00:03.560 | and I've also saw that you've enjoyed the movie
00:00:07.400 | Space Odyssey, 2001, Space Odyssey.
00:00:09.900 | Let me ask the question about Hal 9000.
00:00:12.320 | That makes a few decisions there
00:00:13.860 | that prioritizes the mission over the astronauts.
00:00:17.240 | Do you think, from a high philosophical question,
00:00:20.120 | do you think Hal did the right thing
00:00:22.520 | of prioritizing the mission?
00:00:24.480 | - I think our artificial intelligence
00:00:26.620 | will be smarter in the future.
00:00:28.520 | For a Mars mission, it's a great question,
00:00:30.040 | that is the reality of it.
00:00:31.440 | For a Mars mission, we need fully autonomous systems.
00:00:34.200 | We will get humans, but they have to be fully autonomous,
00:00:37.520 | and that's the most important concept,
00:00:40.840 | because there's not gonna be a mission control on Earth.
00:00:43.640 | 20 minute time lag,
00:00:47.480 | there's just no way you're gonna control it.
00:00:48.680 | So fully autonomous,
00:00:49.760 | so people have to be fully autonomous as well,
00:00:51.500 | but all of our systems as well.
00:00:53.080 | And so that's the big design challenge.
00:00:55.220 | So that's why we test them out on the moon as well,
00:00:57.140 | when we have a, okay, a few second,
00:00:58.800 | three second time lag, you can test them out.
00:01:00.640 | We have to really get autonomous exploration down.
00:01:05.180 | You asked me earlier about Magellan,
00:01:06.680 | and Magellan and his crew, they left, right?
00:01:08.920 | They were autonomous.
00:01:10.400 | They were autonomous, they left,
00:01:12.560 | and they were on their own to figure out that mission.
00:01:14.880 | Then when they hit land, they have resources,
00:01:16.700 | that's in-situ resource utilization,
00:01:18.800 | and everything else they brought with them.
00:01:20.840 | So we have to, I think, have that mindset for exploration.
00:01:23.080 | Again, back to the moon, it's more the testing ground,
00:01:25.080 | the proving ground with technologies,
00:01:27.040 | but when we get to Mars, it's so far away
00:01:29.420 | that we need fully autonomous systems.
00:01:31.440 | So I think that's where, again, AI and autonomy come in,
00:01:35.200 | a really robust autonomy,
00:01:36.500 | things that we don't have today yet.
00:01:38.320 | So they're on the drawing boards,
00:01:39.720 | but we really need to test them out,
00:01:41.160 | 'cause that's what we're up against.
00:01:43.120 | - So fully autonomous, meaning like self-sufficient,
00:01:45.560 | there's still a role for the human in that picture.
00:01:48.160 | Do you think there'll be a time when AI systems,
00:01:51.600 | just beyond doing fully autonomous flight control,
00:01:55.200 | will also help or even take mission decisions like Hal did?
00:01:59.600 | - That's interesting.
00:02:00.440 | It depends, I mean, they're gonna be designed by humans.
00:02:02.480 | I think, as you mentioned, humans are always in the loop.
00:02:04.360 | I mean, we might be on Earth, we might be in orbit on Mars,
00:02:06.720 | maybe the systems, the landers down on the surface of Mars,
00:02:09.880 | but I think we're gonna get, we are right now,
00:02:13.240 | just on Earth-based systems, AI systems
00:02:15.920 | that are incredibly capable,
00:02:18.320 | and training them with all the data that we have now,
00:02:22.520 | petabytes of data from Earth.
00:02:24.560 | What I care about for the autonomy in AI right now,
00:02:27.080 | how we're applying it in research,
00:02:28.840 | is to look at Earth and look at climate systems.
00:02:30.920 | I mean, that's the, it's not for Mars to me today.
00:02:33.240 | Right now, AI is to eyes on Earth,
00:02:35.240 | all of our space data, compiling that using supercomputers,
00:02:39.400 | because we have so much information and knowledge,
00:02:41.680 | and we need to get that into people's hands.
00:02:43.120 | We need, first, there's the educational issue
00:02:45.040 | with climate and our changing climate.
00:02:47.280 | Then we need to change human behavior, that's the biggie.
00:02:50.440 | So this next decade, it's urgent
00:02:52.760 | that we take care of our own spaceship,
00:02:54.200 | which is spaceship Earth.
00:02:55.400 | So that's, to me, where my focus has been for AI systems,
00:03:00.160 | using whatever's out there, kind of imagining, also,
00:03:03.560 | what the future situation is.
00:03:05.200 | What's the satellite imagery of Earth of the future?
00:03:07.600 | If you can hold that in your hands,
00:03:09.300 | that's gonna be really powerful.
00:03:10.520 | Will that help people accelerate positive change for Earth,
00:03:13.960 | and for us to live in balance with Earth?
00:03:15.440 | I hope so, and kind of start with the ocean systems.
00:03:17.760 | So oceans, the land, the air,
00:03:19.600 | and kind of using all the space data.
00:03:21.140 | So it's a huge role for artificial intelligence
00:03:23.920 | to help us analyze, I call it curating the data,
00:03:27.000 | using the data.
00:03:28.240 | It has a lot to do with visualizations as well.
00:03:30.800 | (upbeat music)
00:03:33.380 | (upbeat music)
00:03:35.960 | (upbeat music)
00:03:38.540 | (upbeat music)
00:03:41.120 | (upbeat music)
00:03:43.700 | (upbeat music)
00:03:46.280 | [BLANK_AUDIO]