Page 11 - AIH-1-3

Artificial Intelligence in Health AI in pharma: Embracing transformation

Such fragmentation could fuel negative sentiment if 5.4. Quality control
organizations are overwhelmed by a constant need to adapt Effective oversight will be required to monitor AI so that
to jurisdictional regulations. “red flag signs” can be recognized in a timely manner. In
In a collaborative effort, the Group of Seven (G7) this regard, continual assessments and audits, together with
deliberated international technical standards for AI, the support of XAI, should help preserve the quality of any
60
while the Organization for Economic Cooperation data input, processing, and output procedures. Although
and Development hosted the Global Partnership on AI there is a role for independent regulation, organizations
initiative, aiming to ensure responsible AI development should also take on AI governance to detect and rectify
61
by drawing perspectives from governments, academia, issues associated with their deployed systems. Ultimately,
and industry. The benefits of harmonizing AI regulations regulation could help ease concerns by maintaining
across jurisdictions are evident. Global bodies, such as quality, privacy, and transparency, along with providing
64
the International Council for Harmonization of Technical reassurance about the ability to manage any future risks.
Requirements for Pharmaceuticals for Human Use, could
play pivotal roles in overseeing this process. A cohesive, 5.5. Collaboration
unified stance on regulation could provide clarity for The pace of technological advancement necessitates a
organizations to adhere to best practices when developing collaborative approach. Innovation continues to push
and deploying AI worldwide. the boundaries of knowledge, and as we explore the new
possibilities afforded by AI, it is vital that we collectively
5. Solutions assess the potential societal impacts of any technologies.
5.1. Education Without contemplating the future consequences, we run
the risk of prioritizing advancement at the expense of
As AI applications become increasingly prevalent, it is equity and sustainability. It will be equally important to
imperative that users are informed of their limitations and have effective leadership and communicate a clear strategy
risks. Raising awareness of these aspects can empower in relation to AI, especially considering that investments
users, especially those who are less tech-savvy, to use may yield uncertain returns over the long term.
62
AI knowingly rather than blindly accepting its outputs.
At an organizational level, pharmaceutical companies 6. Innovation
can also take steps to foster a culture of innovation by
promoting a growth mindset, encouraging cross-functional In today’s digital age, innovation has become essential for
collaboration, and investing in the continuous upskilling businesses to maintain their competitive edge, leading to an
and development of its workforce. increased demand within the pharmaceutical industry for
talent with skills in AI. Simultaneously, businesses have
65
5.2. Data protection become more risk-averse, so they may be hindered by
66
AI systems must comply with all relevant data protection financial expenses, time constraints, or a lack of technical
laws and regulations. Risk assessments and safeguarding expertise for innovating with AI. After all, the benefits
measures, such as consent, anonymization and data and compatibility of technology with existing working
67
minimization, should be implemented. Such a combination practices factor into its adoption.
of measures will work to promote security, accountability, 6.1. Quantum cloud computing
and transparency.
Quantum computing holds tremendous promise for AI.
5.3. Transparency Similar to how cloud computing improved performance
In compliance with non-discrimination requirements and efficiency by overcoming the need for companies to own
68
based on protected characteristics, AI algorithms should extensive hardware infrastructure, quantum computing is
63
be trained on diverse, representative data. Such training expected to herald major advancements in processing power
69
processes should be paired with appropriate quality and revolutionize computing. Quantum computing could
assurance checks. This pursuit of inclusivity should also help propel forward the current capabilities of AI in terms
be reflected in the composition of the teams developing of pattern recognition and prediction by enabling data to
69
AI. Explainable AI (XAI) approaches could provide be processed and analyzed at an exponentially faster rate.
much-needed clarity about underlying decision-making In theory, quantum cloud computing would enable
processes, thereby transforming the metaphorical “black rapid innovation at scale, coupling the power of quantum
box” of AI models into a “glass box.” 39(p3506) computing with the scalability of cloud computing.
70

Volume 1 Issue 3 (2024) 5 doi: 10.36922/aih.2973