There is a whole lot of frisson and traction on the recent prognosis of AI exceeding human intelligence. Elon Musk of Tesla has set the benchmark at the end of next year, with the collective intelligence of humans probably being surpassed in the next five years. Contested by Damion Hankejh, Chief executive of Ingk a semiconductor company, who asserts it will not happen and digital machines are not brains. Some have redefined and re-interpreted the objectives of AI. During the 1950’s, Alan Turing in the UK stated that talking to AGI, would be similar to conversing with a person. Steve Wozniak of Apple, has reportedly stated that the AI machine should be able to enter a home and brew a cup of Coffee. Significantly presently, there are no standardised procedures to effectively measure and compare AI.
While previous articles in The Citizen have carried aspects of AI evolution, its scope, business dimensions, as well as some aspects of regulation, we need to peer into its soul. In the public discourse, surprisingly, the focus is on the LLMs (Large Language Models), and Generative AI. There is a secondary focus on the active role of AI, in social and interactive media, as well as its use on the internet and in the newsrooms. What about its use by parties to the conflict, where wars are raging today? AI in these latter spheres is well developed and being advanced constantly, it’s stealthy and shadowy.
Some questions are pertinent. What are the human mental faculties we are concerned about? What exactly is the disruption and dislocation we fear? When focusing on the human MIND, are human capabilities exaggerated? There are many more.
Centuries of evolution have given the Human Brain a framework of functionality. It starts with information or sensory perception. This is processed by a virtual universe of synergistic interactions in the brain, and through this effort, it becomes intelligent thought. Intelligent thought, once again through interaction with external and internal elements can lead to Knowledge. Knowledge over time deliberation as well as interfaces, may result in Wisdom. The survival and evolution of humankind have benefited from precognition. Arguably a unique well-developed faculty in Humans. At its very basic it enabled the visualisation of a hunt and how best to tackle prey, presumably to be successful with no loss of life and limb. To this end, prehistoric cave paintings of hunt scenes are crowded with interpretations. It is felt they also could have been used to actively plan hunts. There is some evidence of spear-notch marks, in support of such wargaming. There are also views that they may have been used to pass on information or knowledge to younger members of the tribe.
Pre-cognition is a powerful quality that enables strategic planning in any domain for an over-the-horizon view. That’s the Human Mind at its best. Looking at a worldview or scenario, in a future time and space. Knowledge, emotions, perception, hunches, imagination, and creativity, all come into play. Above all, the ability to reflect on this, converge and coalesce all this into a comprehensive view and plan. This span of activity seems way out of the reach of AI.
Unfortunately matters in the 21st Century are far from simple. Let’s focus on the first two stages. Information, sensory perception and formulation of an Intelligent View. The bedrock of any human mental progression. The Information Sphere with its several components, has been active and dominant over the last couple of decades. Driven by an attention economy, it seeks to move the mind from Information/incident to wisdom in one sweep. It seeks to eliminate intermediate time for reflection, opinion-making, etc. AI ensures speed, frequency, personalised echo chambers, and an overpowering flood of information. The mass, quantitative assault, repetitive ability and seduction of AI is unmatchable by any human brain. Our brain’s audio-visual segment is approximately 40 %. These senses are put under enormous strain by AI, riding the Information Revolution. It is a struggle to exercise rational thought and reach an objective intelligent view on issues. Let’s not even get into Knowledge and the formulation of strategic outcomes, far less Wisdom. AI has already invaded our personal and collective cognitive spaces. When the appreciation of basic facts and factors is so muddied by effective AI, intellectual activity and LLMs cannot be viewed in isolation.
AI in these mediums has the ability to magnify misinformation, propagate hyper-realistic deep fakes, inundate you with highly personalised content, and attack your intellect multi-dimensionally. We are well aware of high-profile cases such as Cambridge Analytica, what about the everyday assaults that change perceptions and opinions? I cannot resist quoting from a famous Urdu poet, Krishna Bihari Noor, “Sach ghate ya barhe toh sach na rahe; Jhoot ke koi intehan he nahin”- adding or taking away from the truth, renders it not true: for lies there is no test it has to pass. Consequently, curating the truth is disingenuous.
The disruption and dislocation by AI has already manifested. Employee downsizing, selective hiring, hurried retraining and redeployment of the work force are some of the early indications. Significantly there is a lot of ferment in the AI production houses. There is a scramble to lower development costs of the industry, from reducing dependency on Nvidia chips to cheaper locations for power, as well as economical ways to train the LLMs that consume gargantuan quantities of data. This last pursuit has led to the development of Synthetic data, which is orchestrated, cheaper, machine-generated data. To make the LLMs more intelligent, there is a segment of highly qualified computer engineers, whose job is to ask the LLMs the right questions, Prompters. They also aim to reduce the Hallucinations caused by LLMs where completely irrelevant responses and weird answers sometimes manifest. These Prompters, in turn, become irrelevant once the LLM reaches desired levels, and they then have to be reassigned.
It’s one matter to read about the 21st Century workspace and postulate that there will need to be 3/4, career changes, in an average work life. Quite a different matter to live it. Which school or College education the world over is preparing alumni for this kind of employment milieu? It is largely those with both ability and finance, who can hop from ladder to ladder and avoid the snakes. What of the millions of others? Inequity in wealth that is rampant, will soon be complimented by exclusionism in employment. The creation of small islands of selective occupations and wealth, in an ocean of unfulfilled aspirations, is hardly a model for an equitable future.
There is an interesting trend reported in the West, where university enrolments have dropped by up to 12%, in favour of trade schools. Students are seeking trades as avenues of employment, rather than higher education. That’s voting with your feet, for work that may escape AI disruption. Humankind has seen centuries of controlled narratives, propaganda, disinformation and outright deceit. But nothing can compare with the Information Revolution. Its gift was to democratise information and magnify reach, to millions who were previously denied information and Knowledge. The information domain’s marriage with AI, has however manifested painful realities. Profit and business models have given AI a prime handle in Mindplay. We need to recognise, no matter how cognisant we are, the ocean of information and sensory inputs, engulfs us. AI is already on our minds; its future offers both gifts and macabre possibilities.
Lt General Sanjiv Langer PVSM, AVSM retired from the Indian Army, The views expressed here are the writer’s own.