Our future with A.I. – doom or bloom?
The last weeks have been exciting for the tech catalyst team, with one common theme coming up in almost every recent conversation around emergent tech – the impact of Artificial Intelligence on society.
On the 4th of October I spoke at an event called “Breakout: Ethical codes for the digital age”, where here again the impact of A.I. on society was central to the conversation.
I presented two future scenarios that emergent technologies could lead us to; one future whereby technology is applied to enhance society and the environment; and another future where technology is applied in negative ways leading to a future – which certainly I don’t want to live in.
Today we are seeing indications that both scenarios are possible, and of course it is not black and white, but I wanted to deliver a clear message through the demonstration of the two scenarios: that we have the choice of where we want technology to lead, and we have already begun to make these choices.
“It was not the machine, but what one did with the machine, that was its meaning or message.” Marshall McLuhan
Artificial intelligence could create huge opportunities for society, and we have spotted some “weak signals” pointing towards a positive future for tech…
A robot that helps to drive the Circular Economy.
- Drones that monitor emissions of factories and provide automated web reports with data visualization, holding businesses to account more rigorously.
Deep sea robots that monitor the physical and biological background of oceans, forecasting it’s potential for food production and carbon capture.
AI creating resource efficiencies for data centres.
But we can also see “weak signals” of a negative future...
- Weapon-makers are working on an automated gun system that uses artificial intelligence to make "shoot or no shoot" decisions.
- A recent study from Stanford University found that a computer algorithm could correctly identify the sexuality of a person through face recognition. This sparked a debate of whether this kind of software violates peoples’ privacy.
It has been confirmed by the White House that Facebook and Twitter influenced the election results by utilising echo chambers and fake news to their advantage.
Abundant Robotics in California has built an automated apple picker which gives farming efficiency, but at the cost of job losses for low skilled workers.
A software that predicts the defendant’s likelihood of reoffending has already been used by some judges to determine whether an inmate is granted parole or not. This software has been criticised as it may apply additional biases to these cases and lets technology decide the destiny of a person.
There are (legitimate) fears that AI will lead to mass unemployment, as low-skilled job wages drop towards zero, creating an even larger societal gap and bigger inequalities. PWC estimated that 30% of UK jobs are at high risk of being eliminated by AI by 2030
The question is what future do we want to scale up? And how are we going to do it?
A lot of problems occur because emergent technologies develop faster than the regulating systems can respond.
That is why emergent technologies often lack…
# shared ownership
Furthermore an important system dynamic influenced by algorithms is communicated very well by Cathy O’Neil (author of Weapons of Math Destruction):
Algorithms are naturally biased because they are built in order to simplify complex systems. Our real-life biases are transferred to the digital world where they can become even stronger.
O’Neil also describes how biases are often transferred from one system to the other e.g. bias that is applied to a postcode may impact people’s credit analysis or whether a person is invited to a job interview or not…
That means that algorithms stabilize the current systems; opinions, norms, messages and dynamics.
It seems as A.I. as it is currently applied, isn’t a very good system change driver.
However, it will probably be the most influential of all emergent technologies right now, because of its impact potential. It is an overarching emergent technology, like an influencing layer for other emerged technologies e.g. Blockchain, Internet of Things and Automation.
It means that if we channel AI in the right way, we may be able to guide other technologies in the right direction too and have a collective technological impact on societal challenges.
Here are some key questions we have to answer in order to unleash the potential of A.I. for positive change:
• How can we reduce biases of algorithms?
• How do we deal with trade-offs such as job loss especially in low income countries?
• How do we shape A.I. so it helps to drive employment e.g. skills up people and creates new jobs or even helps us redefine wealth and wellbeing?
• How do we define ethical A.I.? Ethics are subjective, so how do we define a shared view on what’s ethical?
• What values do we want to build into A.I.? Where do values clash?
• How can AI be more accountable and algorithm applications more transparent?
How to use systems thinking to channel A.I. into the right direction.
To research and anticipate a technology’s impact seems crucial to avoid unintended consequences and to drive positive impact.
We suggest that the impacts of a technological innovation are considered right at beginning of the process – in the design and innovation process.
The Breakout event allowed us to test a new workshop tool that guides through a process of anticipating different impact levels of a technology.
If you think your business could benefit from one of our technology impact session, please contact me, Michela Rose.
Find further information on 'Tech for System Change'.