AI is the future but who will own all the data?
Autonomous is a word now synonymous with artificial intelligence. It refers to machines that operate free from human direction.
The scare stories abound. Self-driving cars that can’t cope in real life conditions. Result: a pedestrian is mown down and killed in the street.
Or the scandal last month about Cambridge Analytica allegedly using data harvested from Facebook to pinpoint political prejudices and eventually engineer a shock outcome to the US presidential election.
There is a wider definition of autonomous. How can we, as consumers and citizens, remain autonomous as artificial intelligence advances?
Resolving that issue is crucial because artificial intelligence is totally transformative yet needs to be non-destructive.
Take the commonly held but ill-defined view that jobs will be destroyed. A landmark Organisation for Economic Co-operation and Development (OECD) report on artificial intelligence and automation was published last month. The estimates for the number of jobs eliminated by technology range between 47 per cent and just 9 per cent.
The survey found almost half of all jobs would change in nature as a result of new technologies. A smaller number of about 14 per cent were vulnerable to elimination. It further found that manufacturing and agriculture were the most exposed sectors. The threat is also greatest for those entering the workforce and preparing for retirement.
Given that progress appears to be axiomatic, ensuring workers and users thrive is a pressing policy issue. It is partly why the UAE has appointed the world’s first minister for artificial intelligence. Governments around the world are grappling with the challenges and opportunities.
An eye-catching intervention by French President Emmanuel Macron in the sphere last week explored the opportunities and pitfalls. The emerging technology offers great hope for change for mobility, energy, defence, finance and health care.
Mr Macron has made overtures to attract leading companies to his country. The French leader is also savvy enough to register that the advent of the artificial intelligence era is all about rights to data.
He concedes that China and the US are well ahead of Europe. The Chinese government monopolises and exploits data. The US has the innovative technology. US companies that gather data as a result of market-leading positions exploit the material for commercial purposes unfettered.
France and Europe have come up with a third way that puts individual consent at the heart of future progress. It is a very different approach. The continent wants to place ownership of data firmly with the individual. For example, Mr Macron told
Wired magazine that the technology can “totally transform” medical care to make it more predictive and personalised. However, this is necessarily dependent on access to a lot of data.
That is something Mr Macron wants to see but he also wants to avoid opening a Pandora’s box of privacy infringement that hurts the individuals exposed in essentially unknown ways.
While dismissive of warnings of a coming superpower battle or Cold War over the data that underpins the revolution, Mr Macron foresaw sovereignty issues. “How can I guarantee French people that US players will respect our regulation?” he asks.
Europe’s new General Data Protection Regulation (GDPR) comes into force this month and grants sweeping rights to citizens to protect their information. It is, in effect, the opening battle to regulate the use of data by companies.
Many countries outside Europe are looking to ensure their companies are GDPR-compliant. While the US does not have matching laws, some US states such as New York are adopting comparable regulations.
Yet even this imperial approach to regulation could be judged inadequate to the actual needs. The GDPR is strong on consumer rights but has less encompassing proprietary rules for individuals.
There are myriad considerations stemming from the creation of data logs by the machines we use and own. Take, for example, the so-called gig economy where the labour force has few defined rights.
How to transfer the all-important ratings accrued by service providers is an obvious issue of concern. Without entitlement there is a twofold effect.
One is that big companies can exercise oligarchical control, making it hard for challengers to compete on an equal footing.
Second is that the provider has no rights over their own credentials. Having built up a track record on a platform, there is an obvious asset value measure of experience and performance. To lose such a record on departure robs the holder of dynamic credentials in a new business.
Mr Macron went on to warn there is a great danger of social polarisation in AI – that is, that repetitive or “straining” jobs are either eliminated or reduced to menial status. The only winners are the highly skilled and the remaining crumbs are remedial occupations.
The challenge therefore is defined: training people to work alongside the evolving machine world while giving them rights to control their information.
Given that progress is axiomatic, ensuring workers and users thrive is a pressing policy issue