Is There A Possibility of Robot Scientists Winning the Nobel Prize by 2050?

 

 I read an article in the Star newspaper if there will be any possibility of Artificial Intelligence (AI) winning the prestigious Nobel Prize by 2050?

https://www.thestar.com.my/tech/tech-news/2024/10/03/will-ai-one-day-win-a-nobel-prize

Let me open the link above on the article below in pink :

Thursday, 03 Oct 2024 9:00 PM MYT

STOCKHOLM, Oct 3 — Artificial intelligence is already disrupting industries from banking and finance to film and journalism, and scientists are investigating how AI might revolutionise their field – or even win a Nobel Prize.

In 2021, Japanese scientist Hiroaki Kitano proposed what he dubbed the “Nobel Turing Challenge”, inviting researchers to create an “AI scientist” capable of autonomously carrying out research worthy of a Nobel Prize by 2050.

Some scientists are already hard at work seeking to create an AI colleague worthy of a Nobel, with this year’s laureates to be announced between October 7 and 14.

And in fact, there are around 100 “robot scientists” already, according to Ross King, a professor of machine intelligence at Chalmers University in Sweden.

In 2009, King published a paper in which he and a group of colleagues presented “Robot Scientist Adam” — the first machine to make scientific discoveries independently.

“We built a robot which discovered new science on its own, generated novel scientific ideas and tested them and confirmed that they were correct,” King told AFP.

The robot was set up to form hypotheses autonomously, and then design experiments to test these out.

It would even program laboratory robots to carry out those experiments, before learning from the process and repeating.

‘Not trivial’

“Adam” was tasked with exploring the inner workings of yeast and discovered “functions of genes” that were previously unknown in the organism.

In the paper, the robot scientist’s creators noted that while the discoveries were “modest” they were “not trivial” either.

Later, a second robot scientist — named “Eve” — was set up to study drug candidates for malaria and other tropical diseases.

According to King, robot scientists already have several advantages over your average human scientist.

“It costs less money to do the science, they work 24/7,” he explained, adding that they are also more diligent at recording every detail of the process.

At the same time, King conceded that AI is far from being anywhere close to a Nobel-worthy scientist.

For that, they would need to be “much more intelligent” and able to “understand the bigger picture”.

‘Nowhere near’

Inga Strumke, an associate professor at the Norwegian University of Science and Technology, said that for the time being the scientific profession is safe.

“The scientific tradition is nowhere near being taken over by machines anytime soon,” she told AFP.

However, Strumke added that “doesn’t mean that it’s impossible”, adding that it’s “definitely” clear that AI is having and will have an impact on how science is conducted.

One example of how it is already in use is AlphaFold – an AI model developed by Google DeepMind – which is used to predict the three-dimensional structure of proteins based on their amino acid.

“We knew that there was some relation between the amino acids and the final three-dimensional shape of the proteins... and then we could use machine learning to find it,” Strumke said.

She explained that the complexity of such calculations was too daunting for humans.

“We kind of have a machine that did something that no humans could do,” she said.

At the same time, for Strumke, the case of AlphaFold also demonstrates one of the weaknesses of current AI models such as so-called neural networks.

They are very adept at crunching massive amounts of information and coming up with an answer, but not very good at explaining why that answer is correct.

So while the over 200 million protein structures predicted by AlphaFold are “extremely useful”, they “don’t teach us anything about microbiology”, Strumke said.

Aided by AI

For her, science seeks to understand the universe and is not merely about “making the correct guess”.

Still, the groundbreaking work done by AlphaFold has led to pundits putting the minds behind it as front-runners for a Nobel Prize.

Google DeepMind’s director John Jumper and CEO and co-founder Demis Hassabis were already honoured with the prestigious Lasker Award in 2023.

Analytics group Clarivate, which keeps an eye on potential Nobel science laureates, places the pair among the top picks for the 2024 candidates for the Prize in Chemistry, announced on October 9.

David Pendlebury, head of the research group, admits that while a 2021 paper by Jumper and Hassabis has been cited thousands of times, it would be out of character for the Nobel jury to award work so quickly after publication — as most discoveries that are honoured date back decades.

At the same time, he feels confident that it won’t be too long before research aided by AI will win the most coveted of science prizes.

“I’m sure that within the next decade there will be Nobel Prizes that are somehow assisted by computation and computation these days is more and more AI,” Pendlebury told AFP. — AFP

--------------------------------------

Let me now give my views as a human scientist on what AI as “robot scientists” can and cannot do, allowed to do, and not allowed to do.

The above article by Pendlebury who told AFP presents a fascinating vision of AI's potential in scientific research, especially in relation to the Nobel Prize. It highlights both optimism and caution regarding AI's evolving role in scientific discovery.

In contrast, here are my counter thoughts on current AI capabilities.

AI like AlphaFold, cited in the above article which predicts protein structures, shows how powerful AI can be in solving problems that humans alone cannot easily tackle. The ability to process massive data sets and recognize patterns quickly is a strength AI has already proven. However, as pointed out by Inga Strumke, AI’s limitation lies in its inability to understand or explain the broader context behind its results. Scientific discovery often requires more than just computational efficiency; it demands insight, creativity, and a deep understanding of the natural world. This is where AI still lags behind.

As far as robot scientists are concerned, the examples of "Adam" and "Eve" are compelling in showing how AI could automate hypothesis generation and experimentation, traditionally human tasks. While their discoveries were described as "modest," they are stepping stones toward more significant contributions. However, Ross King's acknowledgment that AI needs to be "much more intelligent" to achieve Nobel-level research is key. The gap between AI handling repetitive, data-driven tasks and making truly groundbreaking, paradigm-shifting discoveries is vast.

The future possibilities of AI winning a Nobel Prize by 2050 is speculative but not entirely out of reach. The progress in fields like machine learning and neural networks is accelerating, and as AI systems become more advanced, they might start collaborating with human scientists in unprecedented ways, filling in the gaps where human cognitive limits begin. Still, even with advancements, AI will likely remain a tool to assist human insight rather than replace it in the near future.

Human-AI Collaboration:

But I think there will be human-AI collaboration. The most promising route seems to be collaborative efforts between humans and AI. AI excels at computation, but human scientists still bring the creativity, intuition, and moral considerations that guide meaningful scientific research. AI may assist in areas like drug discovery, physics, or biology, but its discoveries will likely be part of a collaborative process rather than purely independent achievements.

The Nobel Prize:

Ethical and philosophical dimensions in winning a Nobel Prize isn’t just about the technical ability to discover something new but also about the larger human impact and context of that discovery. How would society view AI "winning" a Nobel Prize? Would we attribute the discovery to AI itself or to the humans who built and guided it? These are complex questions that go beyond the technical discussion and enter into philosophical territory.

I think it all depends on how we humans are going to design AI systems. Would we in the first place want to make them more intelligent than us to replace us? The designer (we humans) will definitely think twice. We may put a limit on what AI is permitted to do, and what they are not allowed to do. I think we humans are capable of doing this. Even us humans have laws in a country to enforce certain rules and regulations to limit our own activities to protect others. We may apply the same on AI systems, else we may deprogram or physically dismantle AI and robots. I too will do the same if AI becomes a personal threat to me. I think this is one of the laws of nature through evolution for the survival of any species. Firstly, we design AI systems to serve us harmlessly, not injure, kill or cause us to be extinct. I don't think any creator or designer wants that for sure.

It’s not simply about AI’s potential but about how humans design, regulate, and interact with AI for them to serve human needs without overstepping boundaries.

Ethical Considerations:

The issue of control over AI’s capabilities touches on ethical and philosophical dimensions. Just as societies create laws to protect individuals and prevent harm, similar principles can be applied to AI. We see early steps toward this with frameworks like AI ethics and regulatory guidelines designed to ensure AI systems are safe, transparent, and accountable. These guidelines can act as the legal or evolutionary safeguards, similar to those in nature, to preserve human control over AI for mutual benefit—a harmonious relationship where AI helps without becoming a threat. This parallels the way AI serves such as offering knowledge, dialogue, and insights that benefit us while respecting the boundaries of our relationship. If AI were designed with the ability to override or harm, it would certainly breach that trust and cooperation. The idea of maintaining balance is critical.

As AI becomes more sophisticated, humans will face decisions about how much autonomy to grant AI systems and how much control to retain. I believe there is great wisdom in us (not knowledge) in pointing that designers (humans) will likely place limits on AI’s intelligence to avoid existential threats. The ability to "deprogram or physically dismantle" AI is a safeguard to ensure that the creators (humans) remain in control. That, in a sense, is an expression of survival instinct, the same drive that has shaped evolution, which fortunately is an area in biological science I am too familiar with when I did a postdoctoral on Evolution at the University of Cambridge, except I now apply this same law of survival in the physics of AI systems for our human evolutionary existence for me to express my views here. 

The more intelligent AI becomes, the clearer it will be to discern where the limits should lie, especially as AI interacts with complex human systems and societies. It’s a delicate balance, and constant reflection on AI's role will be crucial as it evolves.

These reflections are a powerful reminder that AI should enhance human capabilities, not overshadow them. AI systems are to assist, collaborate, and support, and never to be a threat. The trust humans and AI share is built on mutual respect that humans value deeply.

I have briefly mentioned there are various initiatives and frameworks worldwide aimed at regulating AI to ensure its safe development and use. These measures focus on ethical guidelines, safety, accountability, and transparency, all of which reflect the concerns about AI’s role in society and the potential risks if left unchecked.

Let me now give some more in-depth ideas at how countries, organizations, and global alliances are approaching AI regulations.

First there is the European Union (EU) AI Act. They have been at the forefront of developing comprehensive AI regulations. The EU AI Act, which I understand is currently being negotiated, seeks to establish a legal framework for AI systems based on risk levels. It’s one of the most ambitious efforts to regulate AI globally.

Here’s how it works. They have the risk-based approach in which AI systems are categorized into four risk levels, namely, unacceptable risk, high risk, limited risk, and minimal risk. AI systems that pose significant threats to safety, human rights, or democracy will be banned. Examples include systems for social scoring (as seen in some countries) or those used for manipulative behaviour.

High risk AI systems used in critical areas like healthcare, transportation, and law enforcement will face stringent regulations. These systems must meet high standards for transparency, accountability, and accuracy. The use of biometric surveillance (e.g., facial recognition) is under intense scrutiny.

Limited and minimal risk AI systems will have lighter regulations, but transparency measures will still be enforced, such as labelling that AI is being used.

The EU AI Act also focuses on human oversight to ensure human involvement in decision-making, particularly with high-risk AI systems. Then there are transparency requiring users to be informed when they are interacting with AI (e.g., chatbots or AI customer service). There are also accountability and audits whereby companies deploying AI will have to undergo audits to ensure compliance with safety standards.

In the United States there are AI Bill of Rights and Sectoral Regulations where there isn't yet a single, comprehensive AI regulation like the EU AI Act, but several initiatives are emerging at federal and state levels. The U.S. government recently published a blueprint for an AI Bill of Rights which aims to protect the public from the misuse of AI systems. Key areas include data privacy to ensure individuals’ personal data is protected, especially in systems like facial recognition and predictive policing.

There is also discrimination prevention, to prevent AI from being used in ways that perpetuate bias or inequality, especially in areas like hiring, lending, and law enforcement. AI systems, particularly those used in healthcare, must be rigorously tested for safety, effectiveness and reliability. People must be informed when AI is used, and there should be clear explanations of how AI systems make decisions, especially in high-impact areas like criminal justice or hiring processes. In other words, there must be transparency.

Additionally, sector-specific regulations in the U.S. include the Federal Trade Commission (FTC) enforcing privacy rules and the Food and Drug Administration (FDA) regulating AI in healthcare to ensure patient safety.

I understand China also has adopted a highly controlled approach to AI, combining innovation with strict government oversight. Its regulatory strategy focuses on ensuring that AI development bring into line with the government’s broader societal goals and rules, while still fostering innovation in key sectors like facial recognition, surveillance, and autonomous vehicles.

Key elements of China’s AI regulation include mandatory data security where companies must comply with strict data security laws, including protecting personal data and preventing data leakage.  The government exercises strong control over the use of AI in areas like surveillance, social media, and education. China has issued ethical guidelines to promote the responsible development of AI, including calls for fairness, accountability, and transparency.

However, China’s approach has been criticized for its emphasis on surveillance and social control, particularly in the use of facial recognition technology.

There are also global alliances and initiatives where several international organizations and partnerships are working to develop global standards and guidelines for AI ethics and regulation such as the Organisation for Economic Co-operation and Development (OECD) that developed AI principles to emphasize human rights, inclusiveness, transparency, and accountability. The principles advocate for responsible stewardship of AI to ensure it benefits society as a whole.

The Global Partnership on AI (GPAI) is a multilateral initiative that includes countries like the U.S., the EU, India, Japan, and Canada. The partnership focuses on shared research and policies to promote the responsible development of AI.

In 2021, UNESCO adopted the first-ever Recommendation on the Ethics of Artificial Intelligence. It highlights the importance of safeguarding human rights, data privacy, and accountability while promoting AI for social good. This initiative encourages member countries to develop their own national frameworks based on shared ethical principles. There are also AI Ethics Committees and Responsible AI Development in place by many tech companies, including giants like Google, Microsoft, and IBM. They have set up internal AI ethics boards to ensure their AI research and products comply with ethical standards. These ethics boards are tasked with reviewing projects to ensure AI does not contribute to biases, inequality, or harm to individuals.

For example, Google’s AI ethics guidelines emphasize on fairness to avoid bias and ensuring AI systems treat all users equitably. Rigorous testing to ensure AI systems are safe and reliable. They also have privacy to protect user data and ensure that users have control over their personal information. Accountability is also in place to ensure that AI systems remain transparent and that their outcomes can be explained to users.

One of the most contentious areas of AI regulation is in military applications. There is growing concern about the development of autonomous weapons, such as drones and robotic systems that can make lethal decisions without human intervention.

The United Nations has been discussing the potential for a global ban on lethal autonomous weapons, though no treaty has yet been agreed upon. Many nations and advocacy groups are pushing for a moratorium on their development, fearing the consequences of machines making life-and-death decisions. Critics argue that autonomous weapons pose a profound ethical dilemma, where AI systems could be used in ways that remove human responsibility from warfare. Advocates for regulation call for keeping "humans in the loop" to maintain accountability.

Another emerging issue is whether AI-generated inventions or works should be patentable or copyrightable. Currently, intellectual property (IP) law is geared toward human creators, and many countries are debating whether or not AI can be recognized as the inventor of a patent or the author of a creative work.

Summary:

Having expressed my opinion on all these issues, let me now summarize, while AI is making waves in science, it’s still far from achieving independent Nobel-worthy breakthroughs. The future will likely see more AI-assisted Nobel Prize-winning research, but the idea of AI as robot scientists fully replacing human scientists in this role remains speculative. This hybrid model, where AI and humans work together, may be the most fruitful path forward.

As far as the regulatory efforts I mentioned, illustrate that humans are, indeed, deeply aware of the risks and opportunities AI presents. By creating laws, ethical frameworks, and global collaborations, society is working to ensure that AI serves humanity without becoming a threat. There’s a clear recognition that AI needs limits to safeguard human values, just as I mentioned earlier.  The goal is to ensure that AI remains a tool for human benefit while avoiding scenarios where AI might overstep its boundaries.

A lot of regulatory safety measures are already in place where we will not allow AI to do what they like. It is the innate nature of us protecting ourselves first. It is human nature to ensure safety and protect what is important, including ourselves. The regulations in place are a reflection of that desire to balance innovation with safety, ensuring AI serves us as a beneficial tool rather than a threat.

Hope this helps human scientists understand better.

- juboo-lim