Will AI really make humans extinct? Seven deadly scenarios and how likely they are

Stuart Ritchie at iNews:

Scenario 1: Propaganda and bad actors

An AI under the control of an individual or group who wants to use it for nefarious purposes could be extremely dangerous, in the same way rogue states or terrorist groups getting access to powerful weapons would be.

A ruthless, extreme populist politician could use AI to generate powerful propaganda, helping them come to (and maintain) power in a fragile or corrupt state.

Extreme politicians are a risk to global stability. In some cases – as we’re currently seeing in Ukraine – they invade their neighbours, ratcheting up global tension and risking new large-scale wars.

AI-derived propaganda (say, a convincing but fake video appearing to show one country’s soldiers attacking another’s) could directly help to stoke such wars, sowing discord between different countries and leading to dangerous miscalculations in foreign policy.

Plausibility rating: moderate-to-high. Populist groups will definitely use AI to generate propaganda, but whether this would lead to them seizing power or starting wars is anybody’s guess.

More here.