The people vs. Amazon: Weighing the risks of its stance on facial recognition and climate change
Amazon’s annual shareholder meeting appears to have highlighted a disconnect between what its staff and senior management think its stance on facial recognition tech and climate change should be.
When in need of a steer on what the business priorities are for any publicly-listed technology supplier, the annual shareholders’ meeting is usually a good place to start.
The topics up for discussion usually provide bystanders with some insight into what issues are top of mind for the board of directors, and the people they represent: the shareholders.
From that point of view, the events that went down at Amazon’s shareholder meeting on Wednesday 22 May 2019 paint a (potentially) eye-opening picture of the retail giant’s future direction of travel.
Of the 11 proposals up for discussion at the event, there were a couple of hot button issues. Both were raised by shareholders, with the backing of the company’s staff.
The first relates to whether Amazon should be selling its facial recognition software, Rekognition, to government agencies. This is on the back of long-held concerns that technologies of this kind could be put to use in harmful ways that infringe on people’s privacy rights and civil liberties.
Shareholders, with the support of 450 Amazon staff, asked meeting participants to vote on whether sales of the software should be temporarily halted. That is until independent verification can be sought that confirms it does not contribute to “actual or potential” violations of people’s private and civil rights.
The second equally as contentious discussion point relates to climate change. Or, more specifically, what Amazon is doing to prepare and prevent it, and drive down its own consumption of fossil fuels. Shareholders (along with 7,600 Amazon employees) asked the company in this proposal to urgently create a public report on the aforementioned topics.
Just say no
And in response Amazon’s shareholders and board of directors voted no. They voted no to a halt on sales of Rekognition. And no to providing shareholders, staff and the wider world with an enhanced view of what it’s doing to tackle climate change.
The company is not expected to publish a complete breakdown of the voting percentages on these issues for a while, so it is impossible to say how close to the 50% threshold these proposals were to winning approval at the meeting.
On the whole, though, the results are not exactly surprising. In the pre-meeting proxy statement, Amazon’s board of directors said it would advise shareholders to reject all of the proposals put up for discussion, not just those pertaining to facial recognition sales and climate change.
Now, imagine for a minute you are a shareholder or you work at Amazon, and you’ve made an effort to make your displeasure about how the company is operating known. And, particularly where the facial recognition proposal is concerned, you get effectively told you are worrying about nothing.
In the proxy statement, where Amazon gives an account as to why it is rejecting the proposal, it says the company has never received a single report about Rekognition being used in a harmful manner, while acknowledging there is potential for any technology – not just facial recognition – to be misused.
Therefore: “We do not believe that the potential for customers to misuse results generated by Amazon Rekognition should prevent us from making that technology available to our customers.”
Commercially motivated votes?
A cynically-minded person might wonder if its stance on this matter is commercially motivated, given two of its biggest competitors, Microsoft and Google, are taking a more measured approach to how and who they market their facial recognition technology to.
Microsoft, for example, turned down a request to sell its technology to a Californian law enforcement agency because it had been predominantly trained on images of white males. Therefore, there is a risk it could lead to women and minority groups being disproportionately targeted when used to carry out facial scans.
Google, meanwhile, publicly declared back in December 2018 that it would not be selling its facial recognition APIs, over concerns about how the technology could be abused.
Incidentally, Amazon’s shareholder concerns about the implications of selling Rekognition to government agencies were recently echoed in an open letter, signed by various silicon valley AI experts. They include representatives from Microsoft, Google, and Facebook, who also called on the firm to halt sales of the technology to that sector too.
Against that backdrop, it could be surmised that Amazon is trying to make the most of the market opportunity for its technology until the regulatory landscape catches up and slows things down? Or, perhaps, it is just supremely confident in the technology’s abilities and lack of bias? Who can say?
Climate change: DENIED
Meanwhile, the company’s very public rejection of its shareholder and staff-backed climate change proposal seems, at least to this outsider, a potentially dicey move.
The company’s defence on this front is that it appreciates the very real threat climate change poses to its operations, the wider world and how its operations may contribute towards that. And, to this end, it has already implemented a number of measures to support its transformation into a carbon neutral, renewably-powered and more environmentally-friendly company as a nod to that.
But, where the general public is concerned, how much awareness is there about that? Amazon has a website where it publishes details of its progress towards becoming a more sustainable entity, sure, but it’s also very publicly just turned down an opportunity to be even more transparent on that.
Climate change is an issue of global importance and one people want to see Amazon stand up and take a lead on, and even some of its own staff don’t think it’s doing a good enough right now. For proof of that, one only has to cast a glance at some of the blog posts the Amazon Employees Climate Justice group have shared on the blogging platform, Medium, of late.
Amazon vs. Google: A tale of two approaches
From a competitive standpoint, its rejection of these proposals could end up being something of an Achilles heel for Amazon when it comes to retaining its lead in the public cloud market in the years to come.
When you speak to Google customers about why they decided to go with them over some of the other runners and riders in the public cloud, the search giant’s stance on sustainability regularly crops up as a reason. Not in every conversation, admittedly, but certainly in enough of them for it to be declared a trend.
From an energy efficiency and environmental standpoint, it has been operating as a carbon neutral company for more than a decade, and claims to be the world’s largest purchaser of renewable energy.
It is also worth noting that Google responded to a staff revolt of its own last year by exiting the race for the highly controversial $10bn cloud contract the US Department of Defense is currently entertaining bids for.
Amazon and Microsoft are the last ones standing in the battle for that deal, after Google dropped out on the grounds it couldn’t square its involvement in the contract with its own corporate stance on ethical AI use.
There is a reference made within the proxy statement that, given the level of staff indignation over some of these issues, the company could run into staff and talent retention problems at a later date.
And that is a genuine threat. For many individuals, working for a company whose beliefs align with your own is very important and when you are working in a fast-growing, competitive industry where skills shortages are apparent, finding somewhere new to work where that is possible isn’t such an impossible dream.
And with more than 8,000 of its employees coming together to campaign on both these issues, that could potentially pave the way for a sizeable brain drain in the future if they don’t see the change in behaviour they want from Amazon in the months and years to come.