Google’s Sundar Pichai doesn’t want you to be clear-eyed about AI’s dangers
Alphabet and Google CEO, Sundar Pichai, is the latest tech giant kingpin to make a public call for AI to be regulated while simultaneously encouraging lawmakers towards a dilute enabling framework that does not put any hard limits on what can be done with AI technologies.
In an op-ed published in today's Financial Times, Pichai makes a headline-grabbing call for artificial intelligence to be regulated. But his pitch injects a suggestive undercurrent that puffs up the risk for humanity of not letting technologists get on with business as usual and apply AI at population-scale - with the Google chief claiming: "AI has the potential to improve billions of lives, and the biggest risk may be failing to do so" - thereby seeking to frame 'no hard limits' as actually the safest option for humanity.
Simultaneously the pitch downplays any negatives that might cloud the greater good that Pichai implies AI will unlock - presenting "potential negative consequences" as simply the inevitable and necessary price of technological progress.
It's all about managing the level of risk, is the leading suggestion, rather than questioning outright whether the use of a hugely risk-laden technology such as facial recognition should actually be viable in a democratic society.
"Internal combustion engines allowed people to travel beyond their own areas but also caused more accidents," Pichai writes, raiding history for a self-serving example while ignoring the vast climate costs of combustion engines (and the resulting threat now posed to the survival of countless species on Earth).
"The internet made it possible to connect with anyone and get information from anywhere, but also easier for misinformation to spread," he goes on. "These lessons teach us that we need to be clear-eyed about what could go wrong."
For "clear-eyed" read: Accepting of the technology-industry's interpretation of 'collateral damage'. (Which, in the case of misinformation and Facebook, appears to run to feeding democracy itself into the ad-targeting meat-grinder.)
Meanwhile, not at all mentioned in Pichai's discussion of AI risks: The concentration of monopoly power that artificial intelligence appears to be very good at supercharging.
Of course it's hardly surprising a tech giant that, in recent years, rebranded an entire research division to 'Google AI' - and has previously been called out by some of its own workforce over a project involving applying AI to military weapons technology - should be lobbying lawmakers to set AI 'limits' that are as dilute and abstract as possible.
The only thing that's better than zero regulation are laws made by useful idiots who've fallen hook, line and sinker for industry-expounded false dichotomies - such as those claiming it's 'innovation or privacy'.
Pichai's intervention also comes at a strategic moment, with US lawmakers eyeing AI regulation and the White House seemingly throwing itself into alignment with tech giants' desires for 'innovation-friendly' rules which make their business easier. (To wit: This month White House CTO Michael Kratsios warned in a Bloomberg op-ed against "preemptive, burdensome or duplicative rules that would needlessly hamper AI innovation and growth".)
The new European Commission, meanwhile, has been sounding a firmer line on both AI and big tech.
It has made tech-driven change a key policy priority, with president Ursula von der Leyen making public noises about reining in tech giants. She has also committed to publish "a coordinated European approach on the human and ethical implications of Artificial Intelligence" within her first 100 days in office. (She took up the post on December 1, 2019 so the clock is ticking.)
Last week a leaked draft of the Commission proposals for pan-EU AI regulation suggest it's leaning towards a relatively light touch approach (albeit, the European version of light touch is considerably more involved and interventionist than anything born in a Trump White House, clearly) - although the paper does float the idea of a temporary ban on the use of facial recognition technology in public places.
The paper notes that such a ban would "safeguard the rights of individuals, in particular against any possible abuse of the technology" - before arguing against such a "far-reaching measure that might hamper the development and uptake of this technology", in favor of relying on provisions in existing EU law (such as the EU data protection framework, GDPR), in addition to relevant tweaks to current product safety and liability laws.
While it's not yet clear which way the Commission will jump on regulating AI, even the lightish-touch version its considering would likely be a lot more onerous than Pichai would like.
In the op-ed he calls for what he couches as "sensible regulation" - aka taking a "proportionate approach, balancing potential harms, especially in high-risk areas, with social opportunities".
For "social opportunities" read: The plentiful 'business opportunities' Google is spying - assuming the hoped for vast additional revenue scale it can get by supercharging expansion of AI-powered services into all sorts of industries and sectors (from health to transportation to everywhere else in between) isn't derailed by hard legal limits on where AI can actually be applied.
"Regulation can provide broad guidance while allowing for tailored implementation in different sectors," Pichai urges, setting out a preference for enabling "principles" and post-application "reviews", to keep the AI spice flowing.
The op-ed only touches very briefly on facial recognition - despite the FT editors choosing to illustrate it with an image of the tech. Here Pichai again seeks to reframe the debate around what is, by nature, an extremely rights-hostile technology - talking only in passing of "nefarious uses" of facial recognition.
Of course this wilfully obfuscates the inherent risks of letting blackbox machines make algorithmic guesses at identity every time a face happens to pass through a public space.
You can't hope to protect people's privacy in such a scenario. Many other rights are also at risk, depending on what else the technology is being used for. So, really, any use of facial recognition is laden with individual and societal risk.
But Pichai is seeking to put blinkers on lawmakers. He doesn't want them to see inherent risks baked into such a potent and powerful technology - pushing them towards only a narrow, ill-intended subset of "nefarious" and "negative" AI uses and "consequences" as being worthy of "real concerns".
And so he returns to banging the drum for "a principled and regulated approach to applying AI" [emphasis ours] - putting the emphasis on regulation that, above all, gives the green light for AI to be applied.
What technologists fear most here is rules that tell them when artificial intelligence absolutely cannot apply.
Ethics and principles are, to a degree, mutable concepts - and ones which the tech giants have become very practiced at claiming as their own, for PR purposes, including by attaching self-styled 'guard-rails' to their own AI operations. (But of course there's no actual legal binds there.)
At the same time data-mining giants like Google are very smooth operators when it comes to gaming existing EU rules around data protection, such as by infesting their user-interfaces with confusing dark patterns that push people to click or swipe their rights away.
But a ban on applying certain types of AI would change the rules of the game. Because it would put society in the driving seat.
Laws that contained at least a moratorium on certain "dangerous" applications of AI - such as facial recognition technology, or autonomous weapons like the drone-based system Google was previously working on - have been called for by some far-sighted regulators.
And a ban would be far harder for platform giants to simply bend to their will.
So for a while I was willing to buy into the whole tech ethics thing but now I'm fully on the side of tech refusal. We need to be teaching refusal.
- Jonathan Senchyne (@jsench) January 16, 2020
A 10-point plan to reboot the data industrial complex for the common good