Wednesday, March 27, 2019

Artificial intelligence and ethics: challenges and responsibilities

Greetings from the Australian National University in Canberra, where Brad Smith, Global President and Chief Legal Officer of Microsoft, is speaking on Artificial intelligence and ethics: challenges and responsibilities. In his introduction, the ANU VC expressed concern over the lack of broadband access around Canberra. He also mentioned the ANU 3AI Institute students who are looking at issues of technology and people (I am sitting in the audience with the students).

This is the second presentation this week on the implications of technology. Yesterday Lieutenant Colonel Keirin Joyce, UAS Sub-Program Manager for the Australian Army talked on "Drones for Good". He handed around a "Black Hornet Nano" advanced military drone, about the size of my thumb. He pointed out that such devices are not "intelligent", they only follow pre-programmed instructions. He also mentioned that drones had been used extensively for disaster relief in the recent Queensland floods.

Brad Smith also was downplaying the current state of AI. He suggested there had been and would not be a sudden achievement of AI. Instead increased computational power and access to data make AI gradually possible. He pointed out some AI is already in routine use, such as in cars for detecting people in the vehicle's path.

Brad Smith raised the issue of ethics with AI and weapons. While suggesting that the laws of war needed to take this into account, he had no specific proposals. I suggest a good start would be for Microsoft to call for China, the United States, and Russia to sign the Ottawa Treaty Banning Anti-Personnel Mines (Australia joined in 1999).

Brad Smith warned of a world like Nineteen Eighty-Four, with routine mass surveillance of the public, preventing free assembly. He proposed laws to limit such surveillance to where there is a court order or an imminent threat. However, Microsoft provides technology which could be used to create a surveillance state. One ethical approach would be for Microsoft to not supply its technology to countries which did not have suitable citizen protections.

Brad Smith ended by raising the issue of a need for a global approach to the issues. One of last slides in the presentation showed statues of Confucius, and Socrates, hinting at the differences of views between China and the West.

Well, I thought that was the end of the talk, but Brad Smith ended on a more positive note, by pointing out three Microsoft initiatives: AI for Earth, AI for Accessibility,  and AI for Humanitarian Action.

No comments:

Post a Comment