MAdtech Point: MAdTech Stack has a far bigger social responsibility in harnessing artificial intelligence

There were two defining moments related technology that came out last week, which has direct impact on the field of advertising and marketing. An artificially intelligent device, billed as the first in the world was launched with built in natural language processing, speech recognition and machine vision. AiPin as it is called, can be controlled by speaking, tapping or projecting a display on the palm of your hand. Potentially this is a “scroller killer”, when done well, this will do what touch screen did to the buttons on our phone. A deepfake video of a leading actress was widely circulated on social media platform, leading to the government issuing an advisory to remove this video failing which the platforms will lose their “safe harbor immunity” status and be liable to criminal and judicial proceedings. Potentially one can never trust any brand promotions/commercials as true unless and until it is now verified. Rashmika Mandanna had the support of the entire industry as the Delhi Commission for Women seeking action, but what about the advertisers and the agencies?

Artificial Intelligence is already at our doorstep, influencing our thoughts and actions. Deemed as the simulation of human intelligence processes by machines, especially computer systems, there are broadly four types of these simulations, three of which are already amidst our lifestyles. At the basic level, called as the “reactive AI”, it is programmed to provide a predictable output based on the input it receives like the various recommendation engines that controls our viewing habit. We have already moved on from this and graduated to using historical and observational data in combination with some pre-programmed information to make some predictions. This “Limited memory AI” learns from the past and builds experiential knowledge like the “news reporter” recently launched by a TV channel.

On the horizon is the “Theory of Mind AI” one that is emotionally intelligent and can hold a meaningful conversation with you, sounding like a real human being. These machines can acquire true decision-making capabilities that are similar to humans with the ability to understand and remember emotions, then adjust behavior based on those emotions as they interact with people. And the ultimate version of the AI is the “Self-aware AI” where machines can be aware of their own emotions, as well as the emotions of others around them, with a level of consciousness and intelligence similar to human beings. Self-aware AI can have desires, needs, and emotions like “The Clock” from Time Variance Authority or “The Entity” from MI: Dead Reckoning.

So far both the Reactive AI and Limited Memory AI has gone unchecked with far reaching implications to the average consumer and the advertising industry. Today we have, artificial intelligence powering recommendation engines and virtual news anchors reading our news reports, while the day is not far off when the “theory of mind AI” starts offering products and solutions as your personal “agony aunt”. I will leave the “Self-aware AI” for a later day to discuss on a different topic and instead focus on the implications and influence of AI in our advertising and marketing technology stack. At the fundamental level, this will all come down to the responsible use of consumer data both as a data fiduciary and as the data principal. As consumers, we have the choice to now give consent to use our data in a responsible manner.

It is advertising that runs the internet. The advertising industry, therefore, has a far bigger responsibility in ensuring that artificial intelligence led solutions does not go unchecked. At the basic level, AI is still dependent on inputs that is provided before it programs a mind of its own at stage 3. It is this “input” that we need to manage well across two basic axes. 1. Who authorised the solution? 2. How authentic is the solution? By definition, input is given to another party other than the one itself, so it all boils down to setting some collaboration guardrails. Collaboration guardrails can only be set when it is not centralised. Logically, decentralisation is possible only on a distributed ledger. Which leads to the question, have we missed the bus on distributed ledger and straight way hopped, skipped and jumped to AI?

The jury is still out, but imagine a situation where all the AI led solutions are built on a distributed platform? By now, we would have known who actually faked the video of Rashmika Mandanna, better still, it would not have happened. We will also have complete control over the engine that recommends our online habits including purchases and entertainment. We will also be in a situation where we will have full visibility to the theories AI can build on its own, as the ownership of this initiative will be federated and centralised.

Artificial Intelligence cannot be centralised and left to certain enterprises only. There needs to be some checks and balances in the form of authorisation and authentication. We need to start this somewhere and perhaps, the advertising industry should take this lead, as after all, it is advertising that is still running the internet.