Google leaves Maven and writes down what it will not do
On 1 June 2018 cloud chief Diane Greene told staff the Project Maven contract would not be renewed. On 7 June Sundar Pichai published principles in which Google undertook not to build AI for weapons or for surveillance violating international norms.
Why it matters
The first time engineers changed who a model is sold to. It is also the first written corporate boundary between AI and the use of force, a document that would be cited for a decade and that the same company would remove in 2025.
The contract became public in March 2018 and The Intercept published its financial detail on 31 May. It was expected to bring in some 250 million dollars a year, against the 9 million figure executives had given. Several thousand employees signed the open letter and nearly a dozen resigned; in May more than 700 academics signed a separate letter demanding termination. Greene announced the decision at a Friday all-hands: work would continue to March 2019 and the eighteen-month contract would not be renewed. In the principles of 7 June, Google listed the applications it would not pursue: technologies that cause or are likely to cause overall harm; "weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people"; technologies that "gather or use information for surveillance violating internationally accepted norms"; and technologies whose purpose contravenes widely accepted principles of international law and human rights. The same document explicitly leaves military work open and names the areas where it would continue: cybersecurity, training, military recruitment, veterans' healthcare, and search and rescue. The Intercept notes separately that Google made no undertaking to forgo other military AI contracts.