Google caves in to pressure, won’t renew Pentagon AI deal

Google caves in to pressure, won’t renew Pentagon AI deal

Google has succumbed to pressure from its employees and announced that it will be exiting from the Project Maven contract with the US Department of Defense next year.

The 18-month contract to apply AI to the analysis of drone footage expires in 2019 and won’t be renewed, Diane Greene, CEO of Google Cloud, told employees on Friday (1 June).

Google’s relationship with the military has been controversial as it plays into the concept of weaponised AI potentially helping to take human lives.

Project Maven seeks to use machine learning and computer vision techniques to improve the gathering of battlefield intelligence from aerial imagery, to help armed forces recognise “objects of interest”.

A chorus of disapproval

Last month, Internet of Business reported that number of employees had resigned from Google in the wake of the deal. Thousands of others signed an internal petition in an effort to persuade CEO Sundar Pichai to withdraw the company from “the business of war”.

Other voices also urged Google to remember its ethical responsibilities to its global user base, famously summed up in its ‘Don’t be evil’ motto.

In April, the Tech Workers Coalition launched a petition demanding that Google cancel the Project Maven contract, insisting that other technology providers also avoid working with the military. “We can no longer ignore our industry’s and our technologies’ harmful biases, large-scale breaches of trust, and lack of ethical safeguards,” the petition read. “These are life and death stakes.”

The world of academia also expressed its concerns over Google’s work with the Pentagon. Last month, over 90 academics in the spheres of ethics, AI, and computer science published an open letter asking Google to back an international treaty prohibiting autonomous weapons systems, and cease work with the US military.

“If ethical action on the part of tech companies requires consideration of who might benefit from a technology and who might be harmed, then we can say with certainty that no topic deserves more sober reflection – no technology has higher stakes – than algorithms meant to target and kill at a distance and without public accountability,” said the letter.

“Google has moved into military work without subjecting itself to public debate or deliberation, either domestically or internationally. While Google regularly decides the future of technology without democratic public engagement, its entry into military technologies casts the problems of private control of information infrastructure into high relief.”

Google has said that it is planning to publish an ethical framework for AI applications this week.

Faustian bargains?

While the likes of Amazon, Microsoft, and IBM are also working closely with the Pentagon, over 30 technology companies – including Facebook and Microsoft, but not Amazon, Apple, or Alphabet  – signed an Accord earlier this year stating that they would refuse to aid any government, including the US, in carrying out cyber attacks.

But money talks for Google/Alphabet and other companies, for whom government contracts are invariably the largest. For example, Google, Amazon Web Services, and Microsoft are still in the running for Pentagon cloud services contracts worth more than $10 billion – known as the Joint Enterprise Defense Infrastructure (JEDI).

Internet of Business says

This morning, the New York Times published emails from earlier this year revealing that Fei-Fei Li, chief AI scientist at Google Cloud, warned the company against using the term “AI” to describe its work on Project Maven.

“Avoid at ALL COSTS any mention or implication of AI,” she said. “Weaponised AI is probably one of the most sensitised topics of AI — if not THE most. This is red meat to the media to find all ways to damage Google.”

Clearly, that assessment was correct, and Google has now caved in to pressure to pull out of the deal – or rather, will not be renewing it. The distinction is important: the 18-month contract still stands, and so Google remains in the business of war.

The urgent advice to avoid linking the term “AI” with Google’s work on Project Maven has a certain irony. In the search-enabled world in which Google is the biggest player – in the West, at least – its chief AI scientist was all too aware that Google’s own algorithms would link the concepts of Google, AI, and warfare in searches, if enough people talked about it.

So it seems that its very dominance in search is perhaps the real reason why it must now withdraw from Project Maven – albeit in slow motion.

Chris Middleton
Chris Middleton is the editor of Internet of Business, and specialises in robotics, AI, the IoT, and technology strategy. He is former editor of Computing, Computer Business Review, and Professional Outsourcing, among others, and is a contributing editor to Diginomica, Computing, and Hack & Craft News. Over the years, he has also written for Computer Weekly, The Guardian, The Times, PC World, I-CIO, V3, and The Inquirer, among many others. He is an acknowledged robotics expert who has appeared on BBC TV and radio, ITN, and Talk Radio, and is probably the only tech journalist in the UK to own a number of humanoid robots, which he hires out to events, exhibitions, universities, and schools.