The Seen Revolution and Unseen After Effects
A Deeper Look into the Capabilities and Responsibilities Surrounding AI
In my last series, "The Invisible Revolution," we embarked on a journey to understand the transformative impact of emerging technologies like AI, robotics, and blockchain. I highlighted advancements in systems such as GPT-2, DALL-E, AlphaGo, and others, arguing that they are setting up the conditions of an unprecedented and often unseen revolution. When it became obvious with the release of Open AI’s ChatGPT in November, the world took notice. Over the course of the past year, its pace of research and development has been nothing short of astonishing. Even for those deeply engaged in its development, it is jaw dropping.
In this article, I wish to further shine a light on the need for AI’s safe development and deployment. There are people who think that AI is merely a prediction machine - that it does what it does by merely predicting the next word in a sentence based on billions of gigabytes of training data. That may have been the case two years ago. I believe these are now more than mere predictive text mechanisms. These technologies are capable of comprehending language, interpreting images, and even strategic reasoning, thanks to deep learning and large datasets. Now, I intend to take a magnifying glass to recent advances that show AI's potential to understand human queries and critically analyse problems. The aim is to advocate for informed leadership and empowered citizens who can guide our collective AI-driven future.
Thanks for reading The Invisible Revolution! Subscribe for free to receive new posts and support my work.
Below is an image of a chemical representation created by a colleague’s sister, who is a post-doc researcher at a leading UK university. Can you try to decipher what it means?
If you were not able to, not to worry. It requires deep specialised knowledge in Chemistry to make sense of it. However, I gave the same prompt to GPT - decipher this image for me. The following is the output that was generated in ~ 10 seconds.
Sure, I'll break down what I understand from this image.
Understanding of the Image:
Fluorescent Nanodiamonds (FNDs) and Cytochrome C (Cyt c): The image seems to depict an interaction between Fluorescent Nanodiamonds (FNDs) and a protein molecule called Cytochrome C (Cyt c). The green structure in the top-left corner represents the protein's three-dimensional conformation.
Molecular Structure: Adjacent to the green structure, there's a detailed molecular structure. This likely represents a portion of the Cyt c molecule or possibly a molecule interacting with Cyt c.
Electron Transfer Process: The arrows and descriptions demonstrate a two-way electron transfer:
From FNDs to Cyt c (labelled "e- transfer from FNDs to Cyt c").
From Cyt c to FNDs (labelled "e- transfer from Cyt c to FNDs").
The NV^0 and NV^+ seem to indicate two states of the nanodiamonds, with NV^0 being the neutral state and NV^+ being the positively charged state.
Electrochemical Test Strip: The strip-like diagram at the center appears to be an electrochemical test strip, showcasing the interaction of Cyt c and FNDs in the presence of an applied voltage, leading to a reduction and oxidation process.
Graphical Analysis: The bottom graph shows a fluctuation in some parameter (likely fluorescence or absorbance) over a series of frame numbers. There are peaks at specific frame numbers, indicating a periodic event – possibly the electron transfer events between Cyt c and FNDs during an applied potential.
Well, pretty impressive but I still did not understand it. So I asked it to give me an analogy related to teaching, hoping that I will be able to relate it with my profession.
Analogy to Teaching:
Think of the electron transfer as a teacher-student interaction. Just as knowledge can flow both ways – from a teacher to a student and vice versa – electrons can flow between FNDs and Cyt c. The peaks in the graph can be likened to 'aha' moments in a classroom, where a concept clicks or understanding deepens.
How do we make sense of this?
While this prowess is impressive, it also underscores the need for ethical development, as unregulated AI poses substantial risks. The leaps in AI capabilities have been significant, moving well beyond the confines of predictive text to instigate shifts in various sectors. However, such power also calls for great responsibility. Our 'Artificial Intelligence' must be governed by ethical guidelines and led by a diverse cohort to avert potential repercussions such as the reinforcement of societal biases or the making of inequitable decisions.
Moreover, it's important to acknowledge that the potential for unregulated AI to alter the course of humanity is profoundly serious. This could manifest as AI-powered warfare or as misinformation campaigns of an unprecedented scale, as witnessed in the conflicts in Ukraine and the Middle East. The seismic effects of such sophisticated misinformation are not abstract; they have the potential to influence our elections and our societal fabric. In this context, there are already deepfake audios and videos featuring prominent figures in scenarios that can stir controversy. It's crucial, therefore, to be vigilant and discerning, recognising that while some deepfakes are easy to spot, others might be wrongfully labeled as such by those with vested interests. Discernment truly is the need of the hour.
Raising awareness, imparting education, and establishing ethical regulation are pivotal. Concurrently, we must not overlook AI’s capacity to address some of humanity’s most pressing challenges, from health crises to climate change. Balancing these scales—between the potential for harm and the promise for good—is one of the foremost tasks of our time."
One of the most transformative potentials of AI that should not be overlooked is its prowess in data analysis. For instance, in healthcare, AI-driven analysis of medical records can identify trends that lead to faster and more accurate diagnoses. In finance, AI algorithms sift through vast amounts of data to predict market fluctuations. However, what’s truly groundbreaking for me is AI’s potential to transform education by analysing patterns in teaching materials and student data to enable personalised learning. AI tutors could identify optimal strategies for different learners and continuously refine instruction based on their engagement and progress. This data-driven approach could help enhance outcomes, promote equity, and make quality education accessible to all.
However, integration requires care to avoid biases and protect privacy. Thoughtful AI implementation could alleviate teachers' burdens by handling rote tasks. But teachers' skills remain vital to nurture students wholistically. With AI assisting - not replacing - motivated educators, education could be enriched for all learners. Keeping humans in the loop is key to developing AI that augments our capabilities and reflects our values. Previously, such deep data analysis was a privilege available to a select few—either due to computational limitations or the expertise required. Now, AI democratises this power, putting sophisticated analytical tools within reach of educators, researchers, and virtually anyone interested in data-driven insights. This is why its development is happening at such a rapid place and we have a catch 22 scenario like never before.
My hypothesis is that dismissing AI as a "dumb predictive text machine" negates its sophisticated capabilities and untapped potential. In the right hands, AI can offer reasoned understanding and valuable synthesis of complex information. Guided by morality, AI can uplift humanity. But without urgent action to embed ethics, we risk profound unintended consequences.
Recent Advances in AI Go Beyond Predictive Text
Remember when AI could only predict your next word as you text? Now, AI systems can:
Examine medical scans to detect cancer or other conditions, saving lives.
Monitor crops using computer vision to alert farmers about disease outbreaks before they spread.
Transcribe doctor-patient conversations to help diagnose illnesses and improve care.
Enable self-driving cars to navigate roads safely using vision, radar and traffic data.
Respond to natural language queries with reasoned analysis
But it is increasingly becoming clear that huge Risks Abound Without Thoughtful Development
While AI holds tremendous potential, its risks are equally formidable:
AI reflects ingrained biases of the limited datasets used to train it, scaling discrimination.
Without transparency, AI makes opaque choices leading to unfair and dangerous outcomes.
Autonomous weapons could initiate warfare at machine speeds without human checks.
Sophisticated AI chatbots can generate fake news and divide societies.
For a technology that can potentially influence careers, disciplines, election outcomes or escalate conflicts, it's crucial to have a robust ethical framework. Below is a snapshot of how AI could impact various disciplines.
Thoughtful frameworks provide philosophies to guide the ethical integration of AI across disciplines. Frameworks like the Bhutan Baccalaureate provide philosophies to guide this ethical integration of AI across disciplines. The Bhutan Baccalaureate, developed by educators in Bhutan, takes a holistic approach to human development using the Five Areas of development as a core framework- Cerebral, Physical, Spiritual, Emotional and Social. In order to develop constructive, contributory citizens of a just and harmonious society, the Bhutan Baccalaureate sees education as a holistic process that extends far beyond the mere acquisition of knowledge. It is a model of education that takes into account a more rounded and wholesome understanding of the learning process.
Just as the Bhutan Baccalaureate seeks to uncover the 'treasures' within each learner, AI should be oriented to unlock the treasures within humanity for societal benefit. The Baccalaureate's multifaceted approach could inform a comprehensive framework for steering AI's impacts across sectors. While AI lacks human consciousness, this model could provide AI developers, designers and policymakers a more holistic ethical checklist.
For example, the cerebral area could focus on transparency, the emotional on building user trust, the physical on safe human-AI interaction, the social on removing biases, and the spiritual on respecting diverse cultural norms. No framework is perfect, but this multidimensional lens could complement existing AI ethics guidelines and help develop innovative, ethical and socially conscious systems. Most crucially, it could help nurture ‘Persons of Substance’ in the AI field who consider diverse ethical perspectives when creating solutions for humanity’s benefit.
The Baccalaureate's focus on "processes", "skills", and "watermarks" provides a model for AI development:
Processes could involve extensive testing, auditing algorithms for bias, and getting diverse feedback to refine AI iteratively.
Skills would include transparency, explainability, and accountability to build user trust in AI.
The watermark imprinted could be one of social responsibility, equity, and human welfare.
Cross-pollination of ideas across disciplines is key. AI researchers should collaborate with experts in ethics, philosophy, social sciences, and humanities to create balanced systems. Just as the Bhutan Baccalaureate nurtures balanced development across five areas, our approach to AI must be guided by diverse perspectives and ethics.
As the waves of an unfolding AI revolution roll in, our compass should not only point towards technological advancement but also wholistic human development. The breathtaking pace at which AI is evolving tends to pull us towards becoming more efficient machines, rather than better-rounded humans. What we need most are Persons of Substance to guide us—individuals whose leadership is born of ethical convictions, who make choices not for profit or popularity, but because they are intrinsically right. These are leaders who inspire rather than command, driven by a vision of human progress that transcends the lure of immediate gains.
It's crucial to remember that we are at the threshold of a new era, one where AI offers more than mere predictions—it promises transformative solutions to humanity's most complex challenges, from healthcare to climate change. While this is cause for optimism, we must temper our enthusiasm with vigilance. The rapid and often opaque development of AI poses risks that can't be ignored, demanding that we approach this powerful technology with a sense of responsibility and ethical commitment.
This ethical layer is all the more critical given the 'black box' nature of AI algorithms. Although issues like security risks and competitive advantage necessitate some level of secrecy, this shouldn't come at the expense of transparency and ethical integrity. We need informed and ethically-conscious leadership across sectors and disciplines, prepared to navigate the complexities and mitigate the risks.
It's also worth pondering how ancient wisdom, encapsulated in texts like the Vedas and Buddhist philosophies, could guide us in this new frontier. These ancient scriptures offer timeless principles of ethical and balanced living that could serve as a moral compass in the development and application of AI. Moreover, these philosophies could inspire us to cultivate an aesthetic mindset that ensures a human touch in AI capabilities, a quality that will become increasingly important as we move forward. As described in the Bhutan Baccalaureate, aesthetics refers to more than just superficial beauty. Rather, it extends to the way individuals operate in the world based on their deeper perception of justice, harmony and responsibility. We must carry this mindset into the realm of AI and shape systems that interact ethically with humanity and the world. This emphasis on holistic aesthetics, on crafting AI that behaves responsibly, will become increasingly important as its capabilities grow more sophisticated. By keeping the human perspective at the centre, we can create AI that uplifts our values rather than undermines them.
AI is an unprecedented form of intelligence, truly alien in many ways—capable yet not conscious, awe-inspiring yet not infallible. Its development and application must be guided by our collective ethical values and societal needs. By taking this balanced approach, we can navigate both the seen revolution and the unseen after-effects, ensuring that AI not only advances our capabilities but also enriches our responsibilities, creating a future that uplifts all of humanity. By taking a balanced approach, we can harness the visible potential of AI while safeguarding against unseen risks.
Thanks for reading The Invisible Revolution! Subscribe for free to receive new posts and support my work.