Artificial Intelligence (AI) has been dominating the headlines for its triumphs, and also fears being expressed by many including some of the best minds in AI. The Association for Computing Machinery released a statement in October 2022 on ‘Principles for Responsible Algorithmic Systems’, a broader class of systems that include AI systems. Several leading AI experts and thinkers have been part of different cautionary messages about AI, issued by the Future of Life Institute, the Association for the Advancement of Artificial Intelligence and the Center for AI Safety. There is deep concern about AI among many who know it. What is behind this?
Areas of use, limitations and AGI
AI systems are capable of exhibiting superhuman performance on specific or “narrow” tasks, which has made it to the news in the field of chess, Go (a game several orders harder than chess) and also in biochemistry for protein folding.
The performance and utility of AI systems improve as the task is narrowed, making them valuable assistants to humans. Speech recognition, translation, and even identifying common objects such as photographs, are just a few tasks that AI systems tackle today, even exceeding human performance in some instances. Their performance and utility degrade on more “general” or ill-defined tasks. They are weak in integrating inferences across situations based on the common sense humans have.
Artificial General Intelligence (AGI) refers to intelligence that is not limited or narrow. Think of it as human “common sense” but absent in AI systems. Common sense will make a human save his life in a life-threatening situation while a robot may remain unmoved. There are no credible efforts towards building AGI yet. Many experts believe AGI will never be achieved by a machine; others believe it could be in the far future.
A big moment for AI was the release of ChatGPT, in November 2022. ChatGPT is a generative AI tool that uses a Large Language Model (LLM) to generate text. LLMs are large artificial neural networks that ingest large amounts of digital text to build a statistical “model”. Several LLMs have been built by Google, Meta, Amazon, and others. ChatGPT’s stunning success in generating flawless paragraphs caught the world’s attention. Writing could now be outsourced to it. Some experts even saw “sparks of AGI” in GPT-4; AGI could emerge from a bigger LLM in the near future.
Other experts refute this vociferously, based on how LLMs work. At the basic level, LLMs merely predict the most probable or relevant word to follow a given sequence of words, based on the learned statistical model. They are just “stochastic parrots,” with no sense of meaning. They famously “hallucinate” facts, confidently (and wrongly) — awarding Nobel prizes generously and conjuring credible citations to non-existent academic papers.
True AGI will be a big deal if and when it arrives. Machines outperform humans in every physical task today and AGI may lead to AI “machines” bettering humans in many intellectual or mental tasks. Bleak scenarios of super-intelligent machines enslaving humans have been imagined. AGI systems could be a superior species created by humans outside of evolution. AGI will indeed be a momentous development that the world must prepare for seriously.
I believe current LLMs and their successors are not even close to AGI. But will AGI arrive some day? I reserve my judgement. However, the hype and panic about LLMs or AI leading directly to human extinction are baseless. The odds of the successors of the current tools “taking over the world” are zero.
Where the dangers lie
Does that mean we can live happily without worrying about the impact of AI? I see three possible types of dangers arising from AI.
Superhuman AI: The danger of a super intelligent AI converting humans to slaves. I do not worry about such a highly unlikely scenario.
Malicious humans with powerful AI: AI tools are relatively easy to build. Even narrow AI tools can cause serious harm when matched with malicious intent. LLMs can generate believable untruths as fake news and create deep mental anguish leading to self-harm. Public opinion can be manipulated to affect democratic elections. AI tools work globally, taking little cognisance of boundaries and barriers. Individual malice can instantly impact the globe. Governments may approve or support such actions against “enemies”. We have no effective defence against malicious human behaviour. Well-meaning people have expressed concern about AI-powered “smart” weapons in the military. Unfortunately, calls for a ban are not effective in such situations. I do not see any easy defence against the malicious use of AI.
Highly capable and inscrutable AI: AI systems will continue to improve and will be employed to assist humans. They may end up harming some sections more than others unintentionally, despite the best intentions of their creators. These systems are created using Machine Learning from data from the world and can perpetuate the shortcomings of the data. They may introduce asymmetric behaviours that go against certain groups. Camera-based face recognition systems have been shown to be more accurate on fair-skinned men than on dark-skinned women. Such unintended and unknown bias can be catastrophic in AI systems that steer autonomous cars and diagnose medical conditions. Privacy is a critical concern as algorithmic systems watch the world constantly. Every person can be tracked always, violating the fundamental right to privacy.
Another worry is about who develops these technologies and how. Most recent advances took place in companies with huge computational, data, and human resources. ChatGPT was developed by OpenAI which began as a non-profit and transformed into a for-profit entity. Other players in the AI game are Google, Meta, Microsoft, and Apple. Commercial entities with no effective public oversight are the centres of action. Do they have the incentive to keep AI systems just?
Everything that affects humans significantly needs public oversight or regulation. AI systems can have a serious, long-lasting negative impact on individuals. Yet, they can be deployed on mass scale instantly with no oversight. How do we bring about effective regulation without stifling creativity? What are the parameters about an AI system that need to be watched carefully and how? There is very little understanding of these issues.
Also read | Learning, thinking, artistic collaboration and other such human endeavours in the age of AI and chatGPT
Many a social media debate rages about AI leading to destruction. Amidst doomsday scenarios, solutions such as banning or pausing research and development in AI — as suggested by many — are neither practical nor effective. They may draw attention away from the serious issues posed by insufficient scrutiny of AI. We need to talk more about the unintentional harm AI may inflict on some or all of humanity. These are solvable, but concerted efforts are needed.
India must be prepared
Awareness and debate on these issues are largely absent in India. The adoption of AI systems is low in the country, but those used are mostly made in the West. We need systematic evaluation of their efficacy and shortcomings in Indian situations. We need to establish mechanisms of checks and balances before large-scale deployment of AI systems. AI holds tremendous potential in different sectors such as public health, agriculture, transportation and governance. As we exploit India’s advantages in them, we need more discussions to make AI systems responsible, fair, and just to our society. The European Union is on the verge of enacting an AI Act that proposes regulations based on a stratification of potential risks. India needs a framework for itself, keeping in mind that regulations have been heavy-handed as well as lax in the past.
P.J. Narayanan is a researcher in computer vision and Professor and (ex-officio) Director of the International Institute of Information Technology (IIIT) Hyderabad. He was the President of the Association for Computing Machinery (ACM) India, and currently serves on the global Technology Policy Council of the ACM. The views expressed are personal
