London, 1st December 2017
Inspired by Russell Buckley’s talk on ‘Exponential Change’, in association with the School of Communications Arts, our Head of Digital Chris Brown gives us a view on the potentially world-changing impact of artificial intelligence and some of the reasons it is not as imminent as we might fear.
The incredible rate of growth in technological advancement in the last 10 years heightened awareness and concern about a future driven, created and regulated by technology.
The field of artificial intelligence, or machine learning, has been one area where progress is outstripping all projections and expectations, leading to fears over the threat that algorithms might pose to life as we know it. From the exhilaration of a world where universal basic income, famine and labour are alleviated by machines, through the annihilation of lethal autonomous weapons systems, increasing financial divide compounded by access to technology and widespread unemployment, expectations are high as to the impact AI will have.
Pondering on this possible future is everyone from American computer scientist and futurist, Ray Kuzweil‘s who believes by 2045 the rise in computer power will overtake humans and AI will run the new Utopic world (aka ‘Singularity Theory’), whilst Tesla founder Elon Musk believes that in order to remain relevant humans must quite literally ’embrace’ tech by becoming cyborgs – so it’s not just companies who need to ‘innovate or die’ it seems…
So, what can we expect in the next ten years? According to Buckley, here’s a few key milestones to watch out for:
- Voice becoming the main way of inputting data (keypads will be defunct)
- Cash will account for less than 1% of transactions (today it’s 5% and Estonia has already abolished cash)
- 50% of workforce will be freelance by 2020 (Deloitte)
- 50% of vehicles will be self-driving (check out BBC Radio 4’s on ‘The Morality of AI‘)
- Security guards will be replaced by drones with guns
- Televisions will stop being sold and broadcast channels will shut down: all content will be on demand
- Para athletes will easily beat fully able athletes. Able athletes will be disabling themselves to run faster
- By 2036, phones will be microscopic (literally)
While all possible predictions, these to me are not because of the advancement of AI, as much as the advancement of our ambition and ability to be creative in applications, but it is full of human ingenuity and input. The fabled and feared AI depicted where our decisions are no longer ours, come from algorithms that think beyond human level intelligence, critically without human input or control.
That a computer will create a computer that’s smarter than a human, and then that new smart(er) computer will create a computer that’s even smarter than the computer smarter than a human… and it goes on.
But is anyone actually doing this? Well, our client Google Deepmind have already created AI that can tech itself to play games more quickly than a human can teach AI to play games, so in a sense it’s already started. And I am sure that one day a computer will develop something we could consider super-intelligence, but I would temper a few of the fears by saying I think it could be a long way off.
While algorithms can already compute, analyse patterns and mine data in milliseconds, and algorithms can learn and teach themselves through reinforcement learning to refine choices, all of these things have tasks programmed by human beings.
We have to define data sets, inputs, value criteria, the field of visibility etc. That’s where it gets super interesting, well at least for me. The AI we see today are not truly creative in the way that we are. The AI that is super-human at processing brain scans for signs of cancer, cannot tell you what song is playing. Not without human intervention. I think the beauty of human intelligence is our dexterity to do multiple things, be spontaneous, be silly, be funny, be imaginative, be sensitive. That to me is true human intelligence.
Right now, we don’t have a clue how the brain works, so the concept we can teach a computer (with all our own input bias) and ask it to create something that can understand something or imagine something that is entirely outside of its input is very hard for me to grasp.
Even when AI developed its own language (http://www.independent.co.uk/life-style/gadgets-and-tech/news/facebook-artificial-intelligence-ai-chatbot-new-language-research-openai-google-a7869706.html) it was still only processing information it had to improve its ability to achieve a set task.
Then on top of that, you have to consider that computer power required to run an algorithm of this enormous data crunching capability is entirely unfeasible. If the human brain were a computer, it could perform 38 thousand trillion operations per second. The world’s most powerful computer, BlueGene, can manage only .002% of that.
If you consider that everything you see through your eyes everyday would fill every data centre we have in the world, you get the size of the problem they face in unregulated learning.
The world of AI and the benefits it can bring us in the future are truly remarkable and the world-leaders like DeepMind, Google Brain, OpenAI, Amazon et al are going to continue to make remarkable breakthroughs in this field, but while I think the much-hyped machine takeover and eradication of humanity is vastly overstated and sensationalised, the need to have the ethical debate and consideration over governance, regulation and control over these systems is absolutely critical to have today.
As with most things, the capability of the technology is vastly in advance of the societal comprehension of its application and the change it will have on people. We need our politicians, academics, policy makers and employers to be actively thinking about these issues, because unlike your iPhone or computer, the consequences for artificial super-intelligence might not have an off-switch.
ABOUT M&CSAATCHI PR’S PARTNERSHIP
WITH SCHOOL OF COMMUNICATIONS ARTS
Since 2015, M&C Saaatchi PR and the School of Communications Arts have formed a unique partnership which has seen the roll out of a new training programme, Intuition, the only one of its kind in the industry. This programme was developed to encourage and improve creativity and to develop softer skills across all of M&C Saatchi PR’s talent. Intuition includes sessions from guest speakers on subjects like ‘Finding Your Creative Flow’, ‘The Chimp Paradox’ and ‘Mindfulness’. Moreover, the CREATE hub of strategists, designers and creatives often collaborate with up-and-coming SCA talent to develop one off creative campaigns that truly speak for and to the next generation.