In the bustling world of AI, new developments are unfolding at a breakneck pace, reminiscent of the historic race to the moon. Prominent organisations such as OpenAI and Google Research are at the forefront, unveiling new models on a continuous basis. This race, fueled by the dream of achieving artificial general intelligence (AGI), harbours both unprecedented opportunities and significant challenges.
The AI sector is a hotbed of innovation, fostering economic growth and opening new avenues for collaboration and knowledge sharing. Initially confined to a small group of experts, the field has now expanded, encouraging a synergy of ideas that could potentially lead to groundbreaking discoveries in AI. This dynamic environment promises the rapid development of new machine learning models that could revolutionise society, provided we navigate the associated risks wisely.
However, this rapid pace of development is not without its perils. It brings to light significant ethical concerns, emphasising the importance of safety in AI development. The field of AI safety engineering is grappling with issues such as AI misalignment, where the goals of AI conflict with human objectives, and the potential misuse of technology for harmful purposes. Moreover, the rapid developments might overshadow these risks, possibly leading to premature deployment of AI technologies driven by short-term incentives.
Socio-economic risks and benefits
These are classic economic dilemmas: On the one hand, the companies want to outpace their competitors, but on the other hand, there is the risk that important core principles (like AI safety and AI alignment) do not receive enough consideration. Safety breaks may slow down the speed of innovation, and some competitors may not prioritise them enough in favour of being faster. Hence, the economy of AI competition has immense upsides (i.e., more innovation and access for the general public), but at the same time, it also has incredible downsides (i.e., the underrepresentation of safety concerns).
To steer this ship safely, global efforts are underway to foster responsible AI development. The European Union, through initiatives in the Horizon 2020 program, is nurturing research excellence and developing socially acceptable machine learning tools. Networks like ELISE and TAILOR are working tirelessly to establish a framework for trustworthy AI, focusing on principles such as lawfulness, ethical considerations, and robustness.
These networks aspire to make Europe a global role model for responsible AI, fostering collaborations with various stakeholders to work towards the realisation of trustworthy AI. However, global regulation will not suffice since it is usually a lot slower than the speed of economic competition. As such, the business players themselves—especially the dominant ones, like OpenAI, Google, Microsoft, or NVIDIA—should form alliances to protect the market and society from overheating, meaning that safety concerns would get drowned in the heat of market competition.
Although such prospects may sound a little frightening to some readers, the reverse may also apply: If market players, educators, and policy makers work together, a bright and innovative future seems to lie ahead. Hence, the business idea of “coopetition” for the sake of all should be prioritised.
Digital humanities
As we delve deeper into the AI era, it becomes increasingly important to maintain a critical perspective on the rapid changes in AI and explore the ethical integration of these technologies in various sectors. The (digital) humanities offer a unique vantage point, providing a critical lens to scrutinise the societal impacts of AI and explore the potential integration of these technologies in humanities research.
AI is poised to become an indispensable tool across all academic disciplines, akin to the universal adoption of computers since the 1940s. This evokes a mandate for “AI literacy,” enabling individuals to collaborate effectively with AI systems without relinquishing critical thinking and decision-making responsibilities. In the humanities, AI has found applications in various tasks, including topic modelling and authorship verification, although its use has sparked controversies regarding methodological correctness and the exacerbation of replication crises in literary studies.
Looking ahead, we envision a symbiotic relationship between humans and AI, where AI serves as a tool providing statistical evidence to support human expertise rather than replacing human judgment Examples may be found in the medical field, where AI may assist in diagnosing cancer more efficiently, as well as advancing biological research by analysing protein structures (e.g., Google’s AlphaFold), thus accelerating discoveries.
Conclusions
As we stand at the cusp of a new era, it is essential to foster a balanced approach where expert insights complement AI outputs, promoting nuanced collaborations between experts and AI systems. This journey through the swift advancements in AI has underscored the necessity for political, legal, economic, and societal interventions to mitigate the associated risks.
The EU’s TAILOR initiative, aiming to foster trustworthy AI, stands as one such example, emphasising the importance of collaboration and raising awareness in this domain. As we navigate this complex landscape, let us embrace AI as a valuable tool, utilising its outputs as evidence while applying our expertise to interpret and frame the results, fostering a collaborative and effective research environment.
Together, let us embark on this exciting journey, navigating the complex dynamics of the evolving AI economy with caution and responsibility, ensuring a future where AI serves as a beacon of innovation and progress, guided by the principles of safety and ethical considerations. Everybody can contribute to this endeavour, regardless of whether they are a consumer, a manager, an educator, a researcher, or a policymaker.
🔬🧫🧪🔍🤓👩🔬🦠🔭📚
Journal reference
Walter, Y. (2023). The rapid competitive economy of machine learning development: a discussion on the social risks and benefits. AI and Ethics, 1-14. https://doi.org/10.1080/07294360.2023.2174083
