Don’t believe ScoMo: What AI means for us in 2020+

Author(s): Dr. Adam Hart

Originally published on Towards AI.

The history of the government of people is superficially a history of the tensions between personal liberty and social responsibility for the greater good. This is what we’re taught about as children and yet I personally believe the major purpose of education is not literacy or numeracy but social conditioning to this mythical golden thread that runs through the polite fabric of societies.

At a deeper level, this tension between liberty and responsibility realizes itself in the western economic model as a race for acquiring capital. Capital in cash and other assets, expertise and now data and performative algorithms. If we have a lot of money we will be free to do what we want.

Similarly to Turing who thought in 1950 that the Turing Test would already be passed, Keynes's view in the 1930s that automation would liberate us from work in 2030 has not come to pass. The debate about labor disenfranchisement due to skill redundancy still plays out.

There are countless articles about automation, the specter of job loss and the 4th Industrial Revolution (4IR). The AI and Robotics community are of course at the forefront of this latest horizon.

While interesting to read because, even if it seems many commentators lack a reasoned theory in which to frame their ‘predictions’, those who do speculate are at least trying to get a handle on the problem, even if they do focus on what new advancement or key invention will be the killer app, the impact of the emergence of AGI, or what innovations in materials science like graphene mean and so forth.

As an alternative, I think that the direction of this kind of speculation is less perceptive than thinking about how AI intersects with the discourses of human nature, economic imperatives, and governmentality. These three discourses are very long-running processes engaged in the question of personal liberty and the centralization of capital in the hands of the few that can provide some insight into where AI may take us.

Based on this perspective I offer up five hypotheses that have a likelihood to play out in 2020+.

Hypothesis A: The shift from technological augmentation to replacement is the area to watch and advise your children about.

Workplace labor disenfranchisement is already well afoot for decades in Australia due to global trends. No car is made in Australia anymore, Toyota robots in Thailand make Australia’s #1 selling car, the Hilux. This is happening irrespective of the Australian Prime Minister Scott Morrison’s (ScoMo’s) idiotic call for children to invest their futures in skill sets that will be augmented and likely replaced in the next few decades on the pretext of a ‘strong economy’.

This inane statement from the leader of a country accompanied by paying a TV celeb $400K AUD for TV advertising to promote the same is like telling your children to be an Uber driver when ubiquitous driverless Ubers are not more than 5 years away and Waymo already have them!

A strong economy is based on the complexity of exports. Australia is the 93rd least complex economy in the World in this regard, out of 100. AI technology is a complex technology that is not being developed much here. This is highlighted by recent concerns that the US is losing the AI ‘arms race’ to China [1].

Whereas it used to be said that Australia is 10 years behind the US in general and scientific advancement, in an ironic inversion, I think Australia is 10 years ahead of the US in labor disenfranchisement trends.

You can learn from Australia’s mistakes.

Hypothesis B: Autocratic leadership of organizations will continue to seek to exploit AI and Robotics to accumulate capital resulting in further erosion of personal freedoms.

As robots become cheaper, and the self-learning DNN’s that power them continue to become used in more generalized activity, and the use-cases for them clearer, it makes business sense to reduce human labor participation. Business doesn’t have to pay a robot, robots don't get sick or require supervision, robots don’t commit fraud, don’t take toilet breaks, can operate 24×7.

We won’t go to work with a trusted team that operates on social equity, we will go to work with an augmented team structure.

While we will attempt to anthropomorphize them, that is incorrect. They are the assets of the owner, who is interested in productivity and efficiency. And they will measure and assess our work.

My son at school has ongoing examinations called ‘continuous assessment tasks’. The AI-enabled agents and Robots will continually assess each human worker’s performance and efficiency.

If today someone can think of and build a workplace toilet that cripples your legs after 5 minutes because it’s tilted at 13 degrees forward, justified on reducing lost working time, then a robot whose job is to monitor your efficiency is more than plausible.

And while populist movies depict robots like C3PO, R2D2 or K-2SO being helpful and friendly assistants, the labor assistants who will become your teammates may not be so helpful to you and your task.

They are assisting the owners' objectives, not yours. Capitalism and efficiency necessarily say it will be like this.

Hypothesis C: Democracy and personal freedoms will slowly give way ‘again’ to 1930’s Fascism.

As I discussed in the Towards AI article The AI ‘shepherds’ that Weaken Democracy, the ubiquitous AI assistants or ‘shepherds’ who know better than us and that surround us will begin to outrank us due to superior data-driven decision making. Due to AI-enabled surveillance, freedom of speech will be restricted, personal liberties constrained. We won’t like it.

Amazon employees and managers are already told-by-code what to do to optimize their work and are apparently make working there a soul-crushing experience. This must only increase as operations research (OR) comes into the fore with in-line AI optimization math that instructs staff on not only when to do this job and by what time but how.

The Japanese led the way here with JIT/Kanban but had a kind of compensatory culture that reduced the inhumanness of it all. The Amazon example shows this approach is not compatible with the western independence ethos. Yet, who would have thought that Hong Kong, that polite financial sector and luxury paradise, would implode into a series of riots because of the restriction of personal liberties?

In the same way, as climate activism is youth-led, students will lead the way for AI-resistance, after all, it is their future.

We’re all taught that if you study and obey the law you’ll have a chance at getting a job and raising a family. As the rules of the game change and there is no hope for some groups, a ‘back-against-the-wall’ mentality may emerge. Groups of disenfranchised humans getting together getting pissed off, giving up on a weak form of democracy that allowed this to happen.

What is happening to democracy today is that leaders are voted in with an expectation set that gets eroded very quickly. Due to the imperative to stay in power, they make odd choices that are not reconcilable with common sense. We lose trust in those we supposedly elected.

Most disastrously, governments will respond in a historical manner to corroborate their own self-fulfilling control structures.

Hypothesis D: Cognitive disorders once thought to have a genetic cause will increase in children due to the inability to distinguish digital from real experiences.

The other day I saw a baby staring at an iPhone in the pram. They would not have been more than 12 months old. I felt so annoyed that the parent had chosen to do this.

With the immersive digital experiences already available and the gamification of workplace experiences forcing us to play games with real things, with Digital Twins and these kinds of tech, with ubiquitous iOS and Android phones being the norm, what is happening is that this advanced tech is pushing its way down the age brackets to childhood developmental stages.

While I am not either a psychologist nor psychiatrist and also have a problem with psychiatric power [2], I understand that these practitioners seek to diagnose cognitive disorders based on an inability to distinguish or respond to different types of stimuli.

As AI-enabled digital experiences become ubiquitous, and especially those using natural language text/speech interfaces that can simulate natural styles, it will likely become increasingly difficult to say if something is real (human) or not (digital).

For the next generation who may not have had the opportunity to grow up without immersive tech, and for who co-working with AI-enabled labor assistants is normal, the ability to distinguish real from non-real is not only more difficult but also perhaps not even a necessary survival skill for employers. But I feel it is a necessary skill to live a good life, remain sane and lessen our ability to be manipulated.

Hypothesis E: Governments will develop increasingly anti-AI retarding legal and regulatory approaches due to declining taxation revenue.

It is most curious, at least in Australia, that I reckon driverless cars will be actively resisted by all state governments. This is for the simple reason that speeding fines and other traffic infringements like not indicating or going through red lights while statistically not resulting in mass harm actually generate an enormous source of taxation for these governments.

The New South Wales government in 2018–2019 earned $4.9B AUD [3] from fines and fees. That’s one state from seven in a small country like Australia.

Driverless cars will be the consummate law abiding errorless drivers.

In this way, state and federal governments will seek to retard AI adoption on the pretext of safety but with a real concern for declining government revenues. Tesla Autopilot is already throttled in Europe.

Labour augmentation leading to replacement will also be retarded. Income Tax and Sales Tax is at stake. Taxation is a life or death matter for all governments. Also, who will buy the goods and services produced by AI-enabled robots if no one is working and earning? This is the circular logic problem that economics can never solve.

For a government to exist, it is both necessary and sufficient that they collect taxation revenue. If they don’t do that, they have no other fundamental purpose and no way to execute on their secondary purposes.

It is ironic given how much people resent tax that taxation alone may give us time to work out how to co-exist with AI-enabled technology and to live a good and sane life amongst an increasingly complex hybrid human/tech eco-system.

Footnote

[1] Although I would note that such language is hyperbole and not helpful.

[2] cf. The Dull and Unpleasant 2020 Ethics of AI-enabled science.

[3] https://www.budget.nsw.gov.au/nsw-budget-2019-20-open-data

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI