Apollo 11: The human factor in autonomous systems

“Once you put someone in orbit, it is a lot harder to fix it if something goes wrong,” says Paul Kostek, a senior Institute of Electrical and Electronics Engineers (IEEE) member and senior systems expert at Base2.

But things do go wrong. About 1:08:53 into the BBC’s excellent documentary, Eight days: To the moon and back, Neil Armstrong and Buzz Aldrin and mission control at the Kennedy Space Center, realise that the toggle for the engine arm circuit breaker on the Lunar Module launch vehicle had broken off. Without it, they would be stranded on the Moon. Nasa engineers worked through the night, investigating how to bypass the switch, but Aldrin figured out he only needed to make sure the electrical contacts in the switch touched, and used a felt tip pen to achieve this, enabling the two astronauts to take off from the Moon’s surface, and ultimately return to Earth safely.

The computers today are far more clever that the ancient hardware that put man on the Moon, but there is a very real sense that the greater the level of automation, the less human operators are kept in the loop, and can understand how to bypass the systems. The tragedy of Lion Air Flight 610 and Ethiopian Airlines Flight 302, are a testament to the fact that automation can make mistakes, and the overrides to enable humans to take back control must be simple and well-understood.