During a walking tour of Queensland’s Daintree rainforest in Australia, a talented guide usually identified creatures that were well camouflaged into their surroundings. At one point, he directed our attention to a tree trunk, where a big grasshopper was camouflaged. The guide’s observations and stories wove together the connections between biology, geology and colonialism, helping explain how big and small changes could transform life on this ecosystem.
(Author), Author provided
Our society has been altered by the rapid proliferation of recent technologies and devices that produce digital data. Nested inside and feeding on this data ecosystem, artificial intelligence (AI) executes cognitive tasks with more potency and speed than human beings. The large-scale transformative power of AI stays camouflaged in plain sight.
Through the lens of the responsible innovation in health research program on the Université de Montréal, we critically examine what lies beyond our immediate experiences of AI.
Artificial intelligence in our lives
Much like driving a automotive, we don’t need to grasp how AI works with a view to use its applications. And just like ways during which the fossil fuel industry shaped the role of cars in our society, AI is delivered through powerful industrial interests and large digital and physical infrastructures. To higher understand their impacts, there’s an urgent must critically appraise how AI delivers its much-touted guarantees.
At the onset of the Industrial Revolution, people in Montréal had no clue in regards to the sorts of infrastructures that were going to be developed to extract, exploit, distribute and use fossil fuels. Montréal was ideally positioned to move goods, including oil, and refineries were later concentrated along the Saint Lawrence River. Beyond negative impacts on residents’ health, the choices made on the turn of the twentieth century to use fossil fuels have had long lasting self-reinforcing effects.
(Wm. Notman & Son/McCord Museum)
And now, within the twenty first century, we’re seeing the changes AI brings and we’d like to think about the wide-ranging ramifications.
Physical support networks
The jewel within the crown of the intangibles economy, AI needs expansive e-infrastructures which have tangible impacts and costs. Estimates suggest “that the carbon footprint of coaching a single AI is as much as 284 tonnes of carbon dioxide equivalent” — five times the lifetime emissions of a mean automotive.
If we decide to use the “oil of the twenty first century,” we can have to construct large powerful computational centres and sizable server farms. AI requires networking and cloud infrastructures to capture, analyze, share and archive vast amounts of information.
When deep learning techniques are involved, training is a key step that consists of feeding the algorithm with large and mostly unstructured datasets. The training of a single AI-based application could also be split over dozens of chips and should require months to finish.
Although it only takes a low energy tap on a smartphone to make use of an application, its development is energy intensive and non-renewable energy sources have a much larger environmental impact.
Energy for training
Thankfully, data scientists are beginning to calculate the energy required to develop AI tools before they’re made available to be used. For instance, a process involved in automating the design of a neural network through trial and error — called the Neural Architecture Search (NAS) — is extremely energy intensive. Without NAS, training the AI tool Transformer takes 84 hours, but with NAS it takes greater than 270,000 hours, thereby “requiring 3,000 times the quantity of energy.”
Reducing the carbon footprint of AI requires a “concerted effort by industry and academia to advertise research of more computationally efficient algorithms” and using more sustainable hardware and model development strategies.
Future policy
Because data generated through digital interactions are price their weight in gold, industrial agreements are prone to keep the long run of AI into the hands of those with corporate interests. Exploiting data to extend corporate profits are the core business of tech giants like Amazon and Google.
This is one in every of the the reason why it will be significant for public policy-makers to create alternative entrepreneurial pathways where data scientists and programmers who aim to design way more meaningful AI can thrive.
Could AI empower those that tackle today’s major societal challenges and seek solutions for the common good? For instance, what would an eco-friendly AI tool to assist us meet the United Nations Sustainable Development Goals seem like? What alternative business and data governance models needs to be promoted for advantages to be shared equitably?
Seeing the forest and the trees could turn a more responsible vision for the twenty first century right into a tangible reality.
This article was originally published at theconversation.com
