- The sheer scale and complexity of AI models makes it ‘impossible to trace their decisions that can hold any individual or military accountable,’ warns Khlaaf, a former systems safety engineer at OpenAI - ‘Amazon, Google and Microsoft are explicitly working with the IDF to develop or allow them to use their technologies … despite being aware of the risks of AI’s low accuracy rates … and how the IDF intends to use their systems for targeting,’ says expert ISTANBUL Israel’s use of artificial