Can AI tell Moral right from Wrong?

As suggested moral choices have much to do with context, and are hard to extract from just words.    I like that changes over time are being considered.  Time is often the most important metadata, and infers the trajectories of choices being made.   We did something similar when seeking to understand the future of marketing messages in context.

Scientists claim they can teach AI to judge ‘right’ from ‘wrong’
The moral machine shows how moral values change over time
Story By  Thomas Macaulay in NextWeb

Scientists claim they can “teach” an AI moral reasoning by training it to extract ideas of right and wrong from texts.

Researchers from Darmstadt University of Technology (DUT) in Germany fed their model books, news, and religious literature so it could learn the associations between different words and sentences. After training the system, they say it adopted the values of the texts.

As the team put it in their research paper:

The resulting model, called the Moral Choice Machine (MCM), calculates the bias score on a sentence level using embeddings of the Universal Sentence Encoder since the moral value of an action to be taken depends on its context.

This allows the system to understand contextual information by analyzing entire sentences rather than specific words. As a result, the AI could work out that it was objectionable to kill living beings, but fine to just kill time.   ... "

Examination of Machine Ethics

In preparation for an upcoming talk and effort that touches on ethics regarding autonomous systems, such as vehicles, but not necessarily restricted to them, I had  reason to look at now the classic 'trolley problem'.   Which is nicely covered in some detail in the Wikipedia entry on The Trolly Problem: 

Implications for autonomous vehicles:
Problems analogous to the trolley problem arise in the design of software to control autonomous cars.[12] Situations could occur in which a potentially fatal collision appears to be unavoidable, but in which choices made by the car's software, such as who or what crash into, can affect the particulars of the deadly outcome. For example, should the software value the safety of the car's occupants more, or less, than that of potential victims outside the car.[33][34][35][36][37]  ... " 

See also the MIT work called the Moral Machine:   https://en.wikipedia.org/wiki/Moral_Machine:

A platform called Moral Machine[38] was created by MIT Media Lab to allow the public to express their opinions on what decisions autonomous vehicles should make in scenarios that use the trolley problem paradigm. Analysis of the data collected through Moral Machine showed broad differences in relative preferences among different countries.[39] Other approaches make use of virtual reality to assess human behavior in experimental settings.[40][41][42][43] However, some argue that the investigation of trolley-type cases is not necessary to address the ethical problem of driverless cars, because the trolley cases have a serious practical limitation. It would need to be top-down plan in order to fit the current approaches of addressing emergencies in artificial intelligence.[44]  ... "