Demystifying Artificial Intelligence

Artificial Intelligence (AI) is the new gift of (IT) that enables the handling of data - created and stored in the binary of computer language - with a speed that surpassed the human capacity for collation and analysis.

Just as IT made human processing ‘smart’, AI makes the technology itself ‘smart’ – being ‘smart’ basically means being able to enhance ‘productivity’ per unit of ‘resource’ whether it is the ‘human’ resource or ‘time’. To the ‘data processing’ capacity is now added the AI’s new achievement of being able to emulate human functions. In this era of multi-media applications, AI is enabling the simulation of voice, ‘looks’ and personal choices to produce authentic looking identities which may lie on the turf of entertainment presently but which should raise alarm for the potential it had of encouraging ‘misinformation’, fraud and social-media misuse for multiple purposes including the game of politics. This identity reproduction is again possible only because technology has the capacity to integrate multiple data fed into the system and collate them for responding to ‘specific commands’.

Of course, the wherewithal for using this kind of AI application is elaborate and expensive - the computer has to have GPU (Graphic Processing Unit) in place of a CPU. AI has surpassed the human capacity to absorb and analyse information but only in quantitative terms - not in terms of the versatility of thought that the human mind is capable of harnessing for examining a situation.

Albert Einstein, the world’s greatest scientist, famously said that ‘imagination is more important than knowledge’– he was alluding not to any wild imagination but to the ability of the human mind to look beyond the facts in front, and to envisage ‘what lay ahead’.

AI essentially serves people in the present without guaranteeing any capacity to control the future - beyond producing some learnings for human beings about what to do or not to do to safeguard the collective good. It is making a huge contribution in the fields of health, education, agriculture, environmental research and above all entertainment.

What has brought AI into the news is the versatile and prolific ChatGPT, an artificial intelligence chatbot developed by OpenAI and released only a few months ago. It is a Generational PRE TRAINED Transform class of language models and is task specific. Its very description, however, confirms the input-output matrix of AI. ChatGPT can have attractive applications. It can mimic the style of celebrities and highly successful business leaders, write business pitches, compose fairy tales and facilitate student essays. Its own acknowledgement of the possibility that incorrect responses could be made, puts an element of risk in its operation.

A fundamental point about the information on the internet is that its reliability should never be presumed. In matters requiring serious decision-making, other corroborating inputs, therefore, should always be sought.

In matters of security, AI may have its use in detecting a breach of a technologically secured perimeter but the real call of security is to have Intelligence - classically defined as information about ‘what lies ahead or what can happen in future’ as regards the plans of an enemy. AI works on data already fed into the system and does not go beyond indicating what would essentially be a repetition made based on collation and interpretation of a combination of patterns already known to it. It can at best throw light on the modus operandi of the adversary as assessed from the latter’s past record without laying any claim on knowing what the current thinking in the enemy camp would be.

AI adds to the threat to cyber security at the levels of nations, business institutions and even individuals. India has plans of addressing these security concerns - under Prime Minister Narendra Modi’s rule, proactive implementation of policy measures has been ensured.

AI studies on machine learning have shown that ‘linear thinking may give way to exponential thinking’ sometimes but this would still be a ‘continuity of the process’ rooted in the analysis of the data already made available. The human mind can use readings of the total environment around an issue as input which is not a given in an AI process. Even the importance of the ‘prompt’ is limited because this trigger essentially lies in what the system has in the present and does not enable AI to respond in the realm that belonged to the future. Predictive findings through AI are useful in a very limited sense in the sphere of security and defence - satellite imagery, signal intelligence and surveillance or communication interception could be helped by it. However, in the fragile geopolitics of today while identification of a ‘threat’ could be refined through the use of highly developed technology symbolised by AI, the nature of ‘response’ must be determined on an exercise of the action taker’s discretion which is a trait given to only a human mind.

In a nuclear setting, however, it is the first strike that has to be determined by the human mind though the second strike could always be automated. This was in fact the deterrence behind the Mutually Assured Destruction (MAD) - the American strategy followed in the Sixties during the height of the Cold War. Nuclear deterrence somewhere still works on the certainty of the far more pervasive second strike.

The best field of AI application, however, is the socio-economic life of the people and it seems the world is already moving towards a consensus on that. It is at present that data analytics and emulation using the ‘Internet of things’ and cloud computing, offer the best utility whether it is distance learning, medical care, diversifying agriculture, aiding business operations or innovating games for entertainment.

New services have made life more purposeful and richer. AI-assisted new models of digital education and human resource development through skilling can help to prepare people for the present knowledge economy. The threat of massive unemployment caused by AI is overestimated since its application will create new services requiring new hands suitably upskilled or re-skilled.

Expansion of technology in the digital domain can create new jobs - Amazon is a good illustration of how the advancement of IT leads to corporate expansion. ‘Deep learning’ AI networks discover intricate structures in the data they handle and learn from the same to produce phenomenal analytics. Industrial use of AI enables plants to optimise their power consumption during live operation and machines to perform quality control checks for making necessary adjustments while manufacturing was still on.

AI has the potential to address some of the biggest challenges in education today through powerful adaptive learning solutions. It has improved crop production and facilitated real-time monitoring including weed detection, yield estimation and crop quality. Domestic use of robots in household chores, such as vacuum cleaning, cutting work in lawns and cleaning the bottom of a swimming pool, can upgrade the quality of day-to-day life. AI in its own ways can prove to be a socio-economic equaliser and work for the amelioration of acute class divides in the society. It is a measure of India’s advancement in digital technology that Prime Minister himself, in his recent meeting with OpenAI’s CEO Sam Altman at Delhi, appreciated AI’s role in enriching the technology eco-system, particularly for the youth of India - Sam Altman later gave out that ChatGPT had been fully embraced in this country.

IT applications are powerful instruments of economic growth, but they are also vulnerable to being misused against India by our adversaries and an awareness of threats to cyber security must therefore run through our sensitive establishments, government organisations and even public bodies engaged in national projects.