00:19It's not like I'm not thinking about like Terminator, I think that they do not have any evil on them.
00:26They just want to complete their task.
00:37If it was the single vehicle on the road, it would work perfectly.
00:42But now we need to interact with a lot of different agents like pedestrian, bicycle, different vehicles.
00:49Each one of them has their own interaction and this is what creates and makes it very complex to test.
00:55Missing because they are not properly testing their AI agents.
00:58The important thing, we are building scenarios.
01:00Reduce the likelihood that something like that will happen.
01:03You can't be 100% guaranteed yet.
01:05Cannot be 100% because you need to keep in mind, especially with agents that are very complex, multiple agents
01:10with tools.
01:11Like you do not have, you still have some blind spots.
01:15Brand for what your brand can do.
01:17So this is your AI agent reading like a…
01:20The agent itself is not the problem.
01:22Like AI is not the problem.
01:23It's obviously the human behind that who is directing and like weighing, like putting weights and let's say directions and
01:31where agents can act or should not act.
01:32So I would actually, again, say the human is accountable for that.
01:36The humans who build the agent…
01:38Again, if we build an agent and an agent does a bad thing, then it should be our responsibility.
01:46These are trusted people and I trust them for short, but there is always commercial factor as well.
01:53Like all the companies are very successful in the Valley, AI related companies.
01:57And would they make the right decision when they are making money?
02:02So I don't know about that.
02:03I think we need to look at our, the core values of the companies that are developing and investing in
02:08AI
02:09and making sure that those core values are being adhered to.
