by Giles Merritt*
“There are the things we know and those we don’t know, and then there’s what we don’t know we don’t know,” confessed a senior American official at the time of the US-led invasion of Iraq. It is truer than ever in the context of setting global rules fit for an Information Age in which we can’t tell where digital technologies are taking us.
If European Union and American negotiators, among many others, are finding the reform of global tax rules a tough nut to crack, as discussed in my June 1 Frankly Speaking article, it’s comparatively straightforward when set against the myriad questions over Artificial Intelligence (AI).
It’s going to be tough to agree on an international code of conduct governing the use of AI and the growing array of digital surveillance technologies. That’s not only because there are so many unknowns but also because these technologies are chiefly deployed by governments that are reluctant to limit their own powers.
Mankind will need to be clear-sighted and cautious when fashioning rules for these technologies
What, then, is known about AI and the computers of the future? Not much by the experts in this rarified field, and precious little by the rest of us. One of the leading think tanks in Washington DC, the Brookings Institution, recently quizzed 1,500 business leaders about the nature of AI and its uses, and a mere 17% of them claimed any familiarity at all.
We know even less about the supercomputers that will power the AI revolution. In the last few weeks a breakthrough has been announced in the US of supercomputing power than can deliver an ‘exaflop’. For the uninitiated, exa is the term used when dealing in quintillions, the definition of which is a billion billions or a million trillions. The supercomputers just around the corner will accomplish a quintillion operations in a single second.
That opens the way to almost unimaginable power. We already know that AI algorithms make their own decisions, independently of any human involvement, on the basis of sensors, digital data and remote inputs. They “act on insights” and are therefore unlike the “passive machines” we know as computers that deliver no more than mechanical and predetermined responses to the questions put to them.
Mankind will need to be clear-sighted and cautious when fashioning rules for these technologies. And the moment to do so is already upon us. Within days of President Biden taking office in January, the Brussels commission presented his administration with a suggested blueprint for transatlantic cooperation on AI.
These technologies have been developing far faster than regulators can negotiate
The EU’s aim is to end the laissez-faire approach favoured by Donald Trump. It advocates replacing the plethora of delegated agencies in the US with a single authority. The EU’s initiative then calls for a risk-based platform of rules that America and Europe would jointly champion as a proposed global body. Facial recognition would be among the technologies slated for much tighter control.
It’s unclear how these rules for AI would fit with existing ones governing personal data protection, nor how to meet the increasingly urgent need for stronger safeguards against cyber-crime, cyber-warfare and the relatively commonplace problems of hacking and ransomware attacks. The EU and the US are far from united on these issues, with China potentially a hugely troublesome outlier with its own very different priorities.
The EU’s insistence on improved data protection has since 2019 seen its Digital Services Act and its Digital Markets Act as unambiguous challenges to the far more open US approach to data processing and transfer. China’s position is at odds with both the US and the EU because Beijing insists on maintaining political control. So far, however, only 14% of the world’s 200 or so national governments have opted to follow the Chinese.
America’s laxer market-based system of oversight has the advantage of stimulating business start-ups and productivity, and is claimed to have won the support of 42% of governments worldwide. The overall winner, though, is the EU with 44% support for its rules-based approach.
What the Information Age’s international rulebook will eventually look like is impossible to tell; these technologies have been developing far faster than regulators can negotiate. But the signs are that in spite of internal tensions, populist pressures and post-pandemic turmoil, the European Union is coming into its own on the shaping of the 21st century’s global economy and the overdue redefinition of globalisation.
*Founder, Friends of Europe
**first published in: www.friendsofeurope.org
By: N. Peter Kramer
