For Country, For Party and For Self: why politicians need to care about AI
The technology offers hope, danger and the potential for victory
Disclaimer: I managed a political party and used machine learning for political purposes. I happen to believe that most politicians lack the time, inclination or motivation to really engage with this technology. That will change.
Why?
For Country:
Globally, societies are increasingly elderly and environmentally challenged. Voters have rising lifestyle and economic expectations. Societies (and the politicians that lead them) must deliver within ever-limited resources.
It is hard to see many tools with the productivity boost potential that AI could provide. Whether maximising utilisation of limited workforces, speeding up key expert and technical bottlenecks (think radiography, weather forecasting…), improving material science engineering, optimising complex processes, piloting remote vehicles or generating personalised educational material AI has a proven role. It is a a general purpose tool that could radically transform the delivery of public services — and foster economic growth.
The Chinese Government understands that AI offers the key to unlocking their demographic trap (“getting old before getting wealthy”), domestic security (and stability) and the face off against legacy US military superiority. Existential issues for the Communist Party of China.
For Western societies the existential trio might be saving the environment, economically sustainable delivery of social services and national security. There are few magic bullets in policy but this is an area where Governments can, and will, do more. And understanding what is possible and what is “AI magic pixie dust” will be key to projecting competence.
For Party:
The debate about AI ethics is increasingly the debate about public ethics. Whether its examining racism in the context of AI bias, educational unfairness unearthed by “mutant algorithms”, child vulnerability and mental health exacerbated by online harms or resource allocation delivered by public service algorithms. To be part of the debate, let alone leading it, politicians need to be immersed in the technology and the conversations coming from it.
Just think of three key public debates that politicians will need to “get”.
- Data privacy versus health and educational outcomes: how should we weight the competing interests of personal privacy versus building data-enabled tools to maximise societal benefits.
- Aggressive personalisation, whether in delivery algorithms, personal pricing and content creation, has huge impact on societal well-being and solidarity. From political schisms to personalised insurance risk or differentiated education journeys these issues are growing.
- Power of the platforms — or more specifically to those who control the data-fed platforms on which we all live and work. And what is national “sovereignty” when they all happen to be owned overseas?
For Self:
Finally — and perhaps most pertinently — politicians will need to understand AI if they want to win elections. Like any old profession politics is not immune to technology. Understanding radio helped give Roosevelt the White House, cracking TV leveraged Kennedy ahead of Nixon, whilst his insight in to the combination of social media and cable news gifted it to Trump.
The first wave of AI in politics was all about maximising message distribution at the lowest possible cost — whether gaming social media algorithms, optimising SEO and online advertising or working out who to target, which doors to knock on with which message.
As the current excitement about Generative AI shows the new wave of AI is about content creation. A decent enough speech can be generated on the fly (as probably can be the answers to most stock media questions). This will not be the Gettysburg address but — lets face it — most political speeches are not that. There are risks — at heart Generative AI is prone to over-confident (and inaccurate) bombast but that is probably true of many an inexperienced speech writer…. check before delivery!
More pertinently the opportunity for personalised messaging now looms. Clearly we do not yet know what will be publicly acceptable and, ultimately, legally allowable but you can be sure that experimentation with both outbound messaging and handling inbound questioning is exploding.
The next decade will see an aggressive expansion of AI tools — as important to our time as the radio, TV and social media have been to previous generations of political campaigners.
