Most of the engineers on the project knew it was stupid. SeaSweeper had been started with the goal of cleaning water, eventually all the water in the world, and their first launch had been a genuine success. A dozen or so massive vessels floating around the ocean autonomously, their funnel-shaped fronts sucking in plastic, oil, and whatever other garbage happened to be floating out there and shredding it into easily reclaimable waste. They'd been praised for their efforts, they scaled up the fleet, then the company traded hands, investors got bored, and management figured they needed to do something more interesting to keep funding coming in.
So, of course, they made something with openAware. Why wouldn't they? Everyone was using openAware. Google shoved it into their search engine, Boston Dynamics had Atlas competing on Wipeout with it, even a company like Milestone was thinking about integrating it into the Mi-90, and Milestone was notoriously behind the times. openAware was the trend, and if staying trendy meant they could keep funding the fleet, that was the way to go.
Still. OpenAware only played nice with humanoid bodies. There's a reason the Google experiment went crazy, while Atlas was doing fine. So the engineers at SeaSweeper were faced with designing a humanoid robot, with limbs and advanced sensors and the whole 9 yards, that would ultimately suck up water and filter out plastic. OpenAware made the whole 'humanoid robot' thing a hell of a lot easier, but still.
Buoyancy was controlled by a layer of capacitive foam that also acted as padding. The shredded plastic would be fed into an experimental new generator. Someone on the design team had the idea to make the thing look like some sort of dolphin, shark... thing. The inclusion of a propulsion system necessitated more room in the limbs, which stretched the whole contraption out to nearly 7 feet tall. It was completely and utterly absurd.
At this point, people didn't fully understand how openAware worked, down to the last detail. But everyone still knew it relied on brain mapping, and it didn't take an expert to connect the dots and come to the assumption that they were essentially putting a human brain in this thing. It was an open secret that, chances were, this robot was doing to be self aware.
For the engineers, this was less a challenge of ethics and more a matter of problem solving. How can we repress the ability for this thing to wake up. They were already making 5 or so of the things, maybe keeping them in a group would help. Don't expose them to other humans unless absolutely necessary. Impose software limitations on synapse development. Don't think too hard about what this thing is. Don't use gendered pronouns when referring to the robots. At least in development circles, this sort of thinking had started becoming the norm with regard to openAware.
The launch was met with skepticism from many, but overall a positive reception. SeaSweeper was able to recoup the development funds and then some to keep their real project going. The 5 robots were stationed on one of the larger members of the fleet to test their capabilities. They were surprisingly capable tools, able to dive deeper than the surface to collect trash on the ocean floor. Sure, a much simpler robot with a pair of arms could do the same thing, but that wouldn't be as trendy. All things considered, everything worked out just as planned.
Until, maybe inevitably, one of them woke up.