Month: January 2019

AI, digital technology, and the future of work

Over recent years a number of concerns about the future of work have been raised. Many concerns focus upon the ‘robots will take our jobs’ slogan. Commentators representing technology firms tend to disagree and argue that many more jobs will be created.

Both sides are right, and we need a strategy to manage the transition to a world of employment dominated by artificial intelligence and digital technology.

The Issues surrounding ‘AI’ and the Future of Employment

• The problem is wider than just ‘AI will take our jobs’, the issue concerns technology in general. Automation does not require AI, merely technology.
• Almost certainly many new jobs will be created as other jobs and aspects of jobs become automated, but the key problem will be the mismatch between the needs of employers and the skills of workers.
• The estimated cost to re-skill workers in the US is $34 billion (simply scaling this to the New Zealand population size indicates an immediate $500 million investment is needed).
• Economic growth requires workers to have jobs (in part because profits accruing to those that own capital tend to accumulate rather than re-enter the economy).
• The developing world will suffer more from automation as it relies on a disproportionate number of manufacturing jobs.
• Whole industries will vanish, so it will be irrelevant whether some are automated or not.
• Many countries are showing slowing of population growth (and the transition to non-sustaining populations). This will have interactive effects with jobs and employment.
• Our present work paradigm links work to income, status and wellbeing. We don’t want to forgo any of these without an alternative strategy.
• Unemployment can lead to stigma and shame. These social ills go beyond loss of income and aren’t easily substituted.
• The value of work goes beyond income, and we don’t want to forgo this value, if we do then even with an abundance of wealth the future is inherently diminished.
• Immigration will not be a sustainable solution in the face of global demand for skilled workers.

A Strategy for the Future

To deal with the concerns listed above, and a host of other associated concerns, we will need a strategy to manage the phasing out of some industries, redundancies as jobs and aspects of jobs are automated, and re-skilling of the workforce to allow transition to growth sectors.

We will also need to care for those workers between jobs and those unable to transition to new kinds of work.

We may further need to:

• Weaken the connection between work and income: UBI has been suggested as a possible strategy, this may require new ways of taxing capital, and new conceptions of what constitutes capital.
• Sever the connection between work and status: we ought to better recognize goodness, kindness, public spiritedness, charity, sustainability and similar traits. By recognizing that many capitalist practices degrade our environment or exploit psychological weaknesses for profit, we can start to recognize that some high ‘status’ individuals are actually anti-community.
• To encourage innovation and novelty we may need to recognize that just as access to water, sunlight, wind and tides is a right for the entrepreneur, access to data/information, processing power, intelligence, and so on are rights in the digital age.
• We need to free up knowledge/IP for common use so innovators can stand on the shoulders of giants and devise solutions to pressing problems.
• We need to curtail the entrenchment of power and growth of inequality because we will need a more equal population for markets to function as intended.
• We may want to protect the right to work (and therefore be productive in the programmes described next).
• Beyond this right, we should move to an obligations based economy, rather than a rights based one. Where companies are obliged to sustain a quota of jobs based on turnover, or fund government programmes of socially valuable jobs/stipends. These may include funding: environmental care programmes, elder and childcare, teaching, the arts, sports, etc).
• If it really is true that as many jobs will be created as are lost, then such programmes will never be needed, so if businesses selling the ‘plethora of future jobs’ dream truly believe their promises, then they will have no qualms about supporting this regulation and consumer confidence will grow.
• If total job numbers do decrease we may need to reduce the working week (and treat work as a scarce resource), perhaps in conjunction with raising the age of superannuation as populations age, allowing us to enjoy leisure time throughout life, rather than all in retirement.
• We will need to provide training and education to aggressively upskill workers. Finland has already taken some steps towards basic AI training for 1% of its population. New Zealand will also need to grow AI (digital) talent.
• What businesses want now will change rapidly so the focus should be on building fundamental capability from the ground up. This will require us to Teach AI and digital skills in schools and to the unemployed.
• To encourage retraining we will need to forgive student debt (and provide low interest loans and free schemes to redo education).
• We need tax breaks for the self-employed who undertake relevant courses part time.
• We need to move to a job market based on skills like interpreting outputs and data, critical thinking, and evaluation thinking (so we can productively work alongside robots).
• We may need a managed decline of population if jobs really do become scarce. This has the added bonus of solving all kinds of other problems, such as climate change.
• We may need urgent research into the economics of degrowth (as opposed to recession, see links to commentary on this approach below).
• We must ensure that worker protections are a trampoline not a safety net. This should entail short term, high investment in those who lose jobs
• We should consider paying users of digital applications for their attention (in an ads based economy).
• We must recognize that experimentation will be required and we must move past conservatism (remembering that stasis helps the oppressors, never the oppressed).
• We need to plan for all of the concerns at the start of this blog, so IF they occur we can respond with the plan we have already devised.

Overall we need to:

– Have ethical debate about the future. We need to decide what we want the future to look like? What would constitute wellbeing? Is it profit? Or work? Is it exponential growth? Or de-growth?

Further reading: The de-growth economy

See Jason Hickel’s blog (a global inequality researcher and Fellow of the Royal Society of Arts)

https://www.jasonhickel.org/blog/2018/10/27/degrowth-a-call-for-radical-abundance

https://www.jasonhickel.org/blog/2018/11/1/a-simple-solution-to-the-growthdegrowth-debate

I want to draw attention to an interview in the Guardian I just read about surveillance capitalism and the allegedly illegitimate conquest of personal data by big digital firms.

The subject of the interview is Harvard Professor Shoshana Zuboff who is author of a book to be released end of January 2019: The Age of Surveillance Capitalism

I’ve previously cited one of Prof Zuboff’s earlier papers in an article about ‘Rapid developments in artificial intelligence: how might the New Zealand government respond’.

I found the Guardian interview so compelling I have already ordered her new book.

The book promises to provide a robust intellectual framework for deconstructing the power of the giant tech firms on the basis of illegitimate conquest of ‘digital natives’ (a very clever metaphor).

This argument could be the forceful rejection of rampant data harvesting that many who oppose the unbridled power of big tech have been seeking.

Here are some key points from the article and interview:

Surveillance capitalism:

• Works by providing free services and enables the providers of those services to gather a phenomenal amount of information about the behaviour of users, frequently without their explicit consent.
• Claims without care for opposing views that human experience is a free raw material that the surveillance capitalist may translate into behavioural data.
• Feeds such data into machine intelligence driven manufacturing processes and fabricates this into prediction products.
• These prediction products are traded in a behavioural futures market.

Why is this a problem?

• The initial appropriation of users’ behavioural data is arrogant. Data are viewed as a free resource, there for the taking.
• The key digital technologies are opaque by design and sustain user ignorance.
• The emergence of these tech giants occurred in a largely law-free context.
• The combination of state surveillance and capitalist surveillance is separating citizens into two groups: the watchers and the watched.
• This is important, because as Jamie Susskind notes in his excellent 2018 book Future Politics, the imbalance in political power has historically been mitigated by the strong being scrutinised publicly and the weak enjoying personal privacy. Upset that dynamic and power shifts.
• Asymmetries of knowledge translate into asymmetries of power.
• We may have some oversight of state surveillance, but we currently have almost no regulatory oversight of surveillance capitalism.

Finally, the key concept that leaped from this interview for me is the following:

• “The idea of ‘data ownership’ is often championed as a solution. But what is the point of owning data that should not exist in the first place?”

Zuboff argues that what we have witnessed over the last two decades is a conquest by declaration. A unilateral decree that some entity may harvest and use a resource freely and without limit.

This is colonial imperialism at its most ruthless and it needs limits.