Throughout its existence, humanity consistently revolutionised its future, from industrial revolution to the current digital revolution. Each iteration of change had a large impact on human cultures and society in general. It seems that we are currently on the precipice of another great revolution, one steered by the power of data and artificial intelligence (AI). It is clear that this will dramatically change the trajectory of human progress. Yet it is unclear by what virtues the changes will come. Will humanity get out-competed by “artificial super intelligence” as proposed by Nick Bostrom? Will capitalistic competition drive AI systems out of our control as proposed by Max Tegmark? Will humanity develop artificial general intelligence in a vacuum? Or will a human-AI symbiotic future prevail? Afterall, we already blur the line between technology and humanity, as for example, technologies such as smartphones can be classified as extensions of ourselves. Following this trend, we propose a future where AI is regarded as an integral element of humanity, rather than an autonomous or detached entity. In its essence, this article takes a stance on the idea that AI systems can and should augment, instead of take over, human decision-making and autonomy. We believe this view can convey a compelling model for the future instead of the fixating on detached human-like or “superintelligent” AI.
“The old idea that man invented tools is a misleading half-truth. It would be more accurate to say that tools invented man” – Charles Clarke
In this article, we will distinguish between artificial narrow intelligence (ANI), artificial general intelligence (AGI), and artificial super intelligence (ASI). All the AI systems that are in use around the world today are considered ANI. Narrow AI systems are built to excel at specific tasks, such as analysing large datasets. An example of which is the AlphaGo system, built by the company DeepMind. This system is able to beat the world champion of the board game Go without a problem but at the same time would get beaten by any 5 year old in a game of chess. To address this shortcoming, many scientists and tech enthusiasts expect that in the future we will be able to develop AGI. As opposed to the specific performance of narrow AI, these systems will be able to solve any task given to them (e.g. those for which they have not been explicitly programmed). Currently, we do not possess the knowledge to build AGI but once we do, these systems will be able to solve any task in a similar fashion as humans (e.g. by reasoning and planning a course of action). Finally, in addition to ANI and AGI, many consider artificial super intelligence (ASI) a possible future. These systems would be truly self-aware and conscious, with intelligence similar to or even outperforming those of humans.
As proposed in the introduction, in this article we will argue for human-AI integration as the next essential step in human progress. Human-AI integration is related to the more common notion of human augmentation, which started as a search for mechanisms to fix the human body and mind which then progressed towards enhancing our cognitive and physical abilities with technology. Human augmentation, therefore, can then be defined as a way to restore or enhance human functions because of integration of artificial systems which have a direct link to human control over them. The most promising technology within this field of study are Brain-computer Interfaces (BCIs) which, in its essence, are devices that analyse brain signals and translate these into commands for other devices that carry out the actions. In recent years, with the adoption of artificial intelligence, advancements have been made in this field. Once these technologies start to mature in the coming years or decades, the future prospects look promising. In the following paragraphs we will elaborate on how this technology can be used to reshape our future.
Space exploration with augmented humans
Humanity’s interest in the final frontier has been daring and grand. We are possessed by exploring the unknown, in turn, this drive expands our horizons in sciences and technology. Not only does this endeavour provide us with a clearer picture of humanity’s place within the universe. It also has provided us with new technologies, such as the GPS, and opened pathways towards new research, improving life on earth. Furthermore, since humanity only lives on one planet, we need to consider existential risks that could affect our only home. Among others, wars with mass destruction capable weapons, climate change and asteroid impacts all have a possibility to snuff out our existence if humanity does not diversify its living conditions. Colonisation of other planets would increase our chances of survival while improving our technology and research capabilities on Earth. Additionally, space has large economic potential as well, asteroid mining could have large implications for Earth’s economy as valuable resources, such as precious metals would provide direct economic growth while making new jobs and contributing to Earth’s sustainability. However, to achieve these milestones, we would require the assistance of AI.
Communion between humanity, AI, and associated machinery is still in its infancy. Nevertheless, much progress has been made. Since the dawn of AI and the golden age of neuroscience, significant strides have been made in brain-machine communication, namely, with the use of BCIs. Recently, innovative methods of fusing AI applications into BCI archetypes have allowed this technology to progress at a much faster rate. With the use of AI techniques, such as machine learning (ML) and convolutional neural networks (CNNs), the BCIs can not only decode neuronal activity of the brain with higher precision but they are also able to handle the ever changing dynamic quality of the human brain. Leading to a more symbiotic relationship between humanity, AI systems and technology. In space, BCIs could be extremely useful in robotics control, cognitive state examination and augmentation. An important avenue is likely to be communication, spacewalks and other astronaut activities require precise and quick communication. BCI technology could become an important aspect in dealing with emergencies as it could be used for extravehicular activities as it is difficult to operate certain nuanced tasks in the spacesuit. BCIs could provide functionality to difficult activities while potentially serving other functions, such as health monitoring. Their practical use, as in the space example, could lead to integration of this technology in other industries, providing new methods of productivity and leisure. Furthermore, similarly to how BCI technology allows amputee patients the control of robotic limbs, likely the first use of BCIs outside of medical purposes would be in space in terms of robotic control. BCI robotic control could be adjusted to be more efficient and less likely to cause logistic problems. The potential future of the use of such BCI technologies for space exploration could lead to dramatic changes in how humans operate both in space and on Earth. Astronauts being closely linked to robots, probes, rovers and other space faring applications could allow them a high degree of control of operations, ability of multitasking, increasing efficiency and safety of space exploration which could likely influence the use of such technologies on Earth. Furthermore, brain-to-brain and brain-to-systems communications could increase productivity and efficiency in unpredictable ways in many human endeavours.
BCI and medicine
In terms of medical applications and health monitoring, BCI technology could increase human survivability in many precarious situations. Currently, invasive BCI technology is already focused on curing blindness and giving movement back to paralysed individuals. In terms of space exploration, it has significant effects on the participants’ health. The launch and landing are especially turbulent. The use of BCIs for the purposes of moment-to-moment health monitoring and remedy administration would be highly beneficial to extreme cases of astronauts and in general use. An example of how this could be vitaly useful are the dangers of radiation in space. Where an integrated AI system would sense radiation levels and use machine learning and neural networks to automate remedy delivery. It is easy to imagine how such a technology could be used for various other purposes. Furthermore, AI techniques and systems can be integrated with humans for remedying cognitive states. For example, machine learning based neurostimulators can help patients with Alzheimer’s disease and Parkinson’s disease. These types of interventions have been largely successful and they show just how technological infusion can be beneficial for humanity. Furthermore, AI methods such as machine learning and deep learning are currently being used as a medium between a blind individual’s eyes and the visual cortex. These remedying applications are likely the tip of the iceberg of what would be possible in the future with human-AI augmentation.
Potential negatives of Human-AI augmentation
Given the fact that these AI based implants and their other applications are likely to be implemented in the future to various public, private and commercial industries, it is important to consider the possible dangers and drawbacks of their use. The more obvious, yet a major concern is the possibility of implants being hacked, leading to decrease in human autonomy, and in a potential dystopian future: mind control and manipulation. In our time, implants already see the light of day, as an American company provided an option for its employees to be infused by microchips into their hands for access to the premises and computers. However, if these technologies are going to be implemented within our society, then policies and laws would need to be caught up and adequately prepared to make these changes safe and secure. In terms of BCIs, they often are directly involved in deciphering and mapping neuronal patterns and activities. Therefore, their use raises some important questions with regards to autonomy and privacy as their ability to assess individuals behaviour and all the caveats that accompany it could entail manipulation of people’s behaviour in various contexts, such as political affiliation and moral decision making.
Besides the direct impact integrated AI technologies can have, major societal transformations need to be weighed up, do the perks associated with the integration outweigh the potential pessimistic consequences? It is difficult to predict what changes such deep integrations will bring to our society. In many ways, AI augmentation of our society is a step into mystery, there are likely to be many unforeseen consequences. On a majorly pessimistic side, augmentation tech such as BCIs could be used for marginalising or causing violence against certain groups in society, as we can see in the Uyghur’s situation. It would truly be wasted potential if augmentation technology would be used for surveillance and control. Of course, this could also be extrapolated to bleak use of surveillance in a capitalist sense as well, with targeted advertisement, election tampering or overtaking governments in terms of influence and power. Furthermore, there has been a significant amount of query done on the impact of such technologies on the future of employment, mostly implying that such impactful technologies would limit menial and routine employment. Lastly, it could be argued that augmentation would segregate the poorer populations from obtaining the technological benefits.
Potential solutions of Human-AI augmentation
However, human augmentation can be seen as progress for humanity and of great benefit to individual users as well. Since the future is unknown, the aforementioned pessimistic scenarios could likely be untrue and the tech would be used to counter segregation, inequality and despotism. This is in hope that humanity learns from its previous historic mistakes of abusing technology. In terms of societal impacts, there is no guarantee that future governments would be control focused dictatorships, instead, augmentation could be used strictly for anti-terrorism functions or providing rapid medical assistance. In terms of surveillance capitalism, in recent years there has been an inclination for large corporations to work with governments and a minor reduction in purely profit driven nature. If this trend continues, the likelihood of misuse of human augmentation would likely decrease. The argument that many people of different backgrounds would lose employment is not necessarily true if one considers the history of technological advancement. Humanity has endured many technological revolutions, from industrial revolution to atomic power, mankind adapted and thrived with these changes and in turn other types of employment emerged. Lastly, for the aforementioned social segregation of technologies argument it could be challenged by how previous technologies were adapted, for example, the use of mobile phones was initially adapted by a select few. However, nowadays, a large majority of humanity uses smartphones. Therefore, elevating humanity’s capabilities past our evolutionary limit could have both negative and positive consequences, but every advancement happened with both possibilities and our experience over the last six million years demonstrates that technology can be used for both betterment and harm. This means that we can use our history and knowledge to steer human augmentation towards betterment and hope for the best.
Why artificial super intelligence is NOT the next essential step
As opposed to human augmentation, humanity could also focus its trajectory towards the creation of conscious AI. Many people envision ASI as the next essential step in human progress. One of the advocates of ASI, or conscious AI technology, is Kanai, who argues that conscious AI will have a lot of benefits for humanity. In his article, he mentions some of these benefits. One important aspect of conscious AI he claims, is the ability of these systems to provide insight into their workings:
“If we can’t figure out why AIs do what they do, why don’t we ask them?” – Kanai
The problem with this reasoning, however, is that this view seems to assume that if we are able to create a consciousness in artificial entities, it would be like communicating to a human who has complete control and knowledge over all of a computer’s abilities. This line of reasoning assumes, for example, that conscious AI systems will have the ability to tell us how their inner mechanisms work. Meanwhile, the average human would be completely in the dark when asked to explain how his or her eyes, combined with their brain, process lightwaves so that they can have an understanding of the outer world. In other words, just because we are conscious does not mean that we have any clue about what is going on inside us. Why then, do we make the assumption that ASI systems would be able to tell us exactly what is going on inside them and can control everything that a computer has to offer. It is more likely that if we are ever able to create a consciousness inside a system, we still have to develop the mechanisms for this consciousness to be able to control all aspects of its environment. One could ask then, why would we try to develop an artificial consciousness in the first place? If we still have to develop the means for such a system to control its environment, then we should also be able to use this knowledge, combined with the aforementioned BCI technology, for a human consciousness to stay in control.
The aforementioned assumption is just one of the many assumptions we as humans make when we think about the future of ASI systems, as all our ideas about these systems are rooted in imagination as opposed to scientific evidence about the workings of such systems. Which makes sense because we have no clear trajectory towards this technology. This becomes evident when we consider that the discussions relating to conscious AI are only meaningful if we are able to develop an artificial consciousness in the first place. Kanai argues that if we consider introspection and imagination as two important factors of consciousness, the creation of conscious AI is inevident and that this should be the focus of our research. However, as the debate about artificial consciousness shows, the scientific community is still very much divided into how the human consciousness emerges and how we would go about developing this in an artificial entity. As Gee Wah and Wang Chi point out:
“Thus far, nobody has any idea exactly what consciousness is in humans. Thus, that brings us back to the basic problem, i.e. we do not have the know-how to design the AI system with the same consciousness as humans, since we do not know it ourselves.” – Gee Wah and Wang Chi
Some even argue that it is impossible to compute self awareness because consciousness is different from matter, which would mean that we will never be able to create ASI.
It is clear that we currently have no idea on how to create an artificial entity with a consciousness and, therefore, there is no reason to assume that we will be able to develop this anywhere in the near future without a scientific breakthrough. We would argue it is illogical to pursue a trajectory that is rooted in a random scientific breakthrough to happen as, again, the evidence shows there is no clear path towards the creation of ASI. Wouldn’t it, then, make more sense to invest our resources and energy in the pursuit of human-AI integration, as this technology is already being implemented as we speak and there is a clear trajectory towards its future?
Even if we forget about the practicalities that make it unlikely for ASI technology to develop in the near future, we should ask ourselves if this is what we want for our future. One quick search on artificial super intelligence or conscious artificial intelligence shows vast amounts of articles mentioning the potential dangers of this technology, with many considering it an existential threat to humanity as we would lose our position as the dominant species on this planet. A common conception among these articles is that ASI systems will be impossible to control because we can’t begin to imagine the possible ways in which such a system might mark the end of humanity. Considering all of this, it is much more reasonable to envision a future of human augmentation as there is a clear trajectory towards this future and we as humans remain in control.
To conclude, in this article we proposed the view that the next essential step in human progress is human-AI integration. The current state of artificial intelligence research can be considered as if we are at a crossroads where the direction we decide to take will define our future. We would argue that if the collective fear is that humanity will be surpassed by artificial systems and made obsolete, making AI an extension of ourselves and using it as a tool for human augmentation in various aforementioned applications is the best path forward to secure our future on Earth and outside of it. The practical applications of integrated AI technologies are likely to encourage the use of such tech in other sectors and eventually, in everyday life, changing the trajectory of human development forever.