Making AI Trustworthy: Navigating the future of innovation and ethics

Image: Shutterstock

Generative AI is a remarkable tool, a technological marvel that wields the power to craft realistic and imaginative content. Its influence has seamlessly intertwined itself into the very fabric of modern society, propelling us towards chartered realms of progress and reshaping the world as we perceive it. With boundless potential, generative AI offers organisations a gateway to innovation, fostering economic growth and promoting societal well-being.

Generative AI is a groundbreaking force that transcends boundaries, a force that breathes life into intelligent applications capable of understanding and responding to human language with great finesse.

Also Read: Generative AI could add up to $4.4 trillion a year to global economy: McKinsey

Unlocking the Potential: How Trustworthy AI Reshapes Industries

In the business realm, there is growing interest in how generative AI can revolutionise enterprises. In the pharmaceutical domain, Generative AI has already become a formidable ally in the quest for drug discovery and development. Likewise, in the fast-moving consumer goods (FMCG) industry, AI breathes life into new products and crafts marketing slogans that resonate with consumer preferences and trends.

However, as generative AI expands its reach, it reveals a landscape filled with strategic complexities, financial challenges, and the unrelenting scrutiny of public perception. The trustworthiness of generative AI hinges on how organisations utilise it. As enterprises venture into this rapidly evolving AI domain, trust and ethics factors should be considered and addressed.

Also Read: Is Generative AI a disruptive paradigm-shifter? The questions to ask

Risks and Ethical Considerations: The Unchartered Terrain 

Many risks and ethical considerations emerge as the inevitable shadows of its brilliance in this burgeoning frontier of Generative AI. Some of the risks and limitations associated with Generative AI include:

1) Hallucination: A scenario when a false/out-of-context response is generated to a query. This response could be grammatically or sometimes even semantically incorrect. This is due to pre-trained results.

2) Token Size Limits: The maximum token size limit per model has increased. It is 16K (GPT 3.5 Turbo) in a single call, which can make it more challenging to process larger documents.

3) Adversarial Behaviour: It is critical to minimise risk proactively, which can be from malicious behaviour on the network. This is necessary to maintain operations and customer trust.

4) Ethical Use: Organisations must ensure that AI usage aligns with the purpose of the overall exercise, and it is imperative to assess manual intervention in the loop when evaluating AI's suggestions.  

Moreover, generative AI systems harbour the potential to fabricate sensitive personal data, such as medical records and financial information. The exploitation of this data could fuel identity theft and other criminal activities. Thus, the onus lies on stakeholders to diligently assess these risks and proactively mitigate them.

Addressing the Risks

India's Newly Enacted Data Privacy Law

India's Digital Personal Data Protection Act 2023 is a new law regulating how individuals' digital personal data is collected, processed, stored, transferred, and deleted by various entities and individuals. This legislation stands poised to safeguard citizens' fundamental right to privacy, extending its purview across real and virtual domains.

Also Read: 'With generative AI, spatial storytelling is at an inflection point'

However, this newly instituted law brings generative artificial intelligence (AI) models into potential conflict. The legislation mandates the protection of personal data by platforms and necessitates explicit user consent for ongoing data processing. Consequently, organisations utilising generative AI to develop AI applications must ensure stringent compliance with the law.

Charting the Path Forward

The imperative for trustworthy AI frameworks intensifies as the prominence of Generative AI continues to ascend. To navigate this evolving landscape successfully, several pivotal factors warrant incorporation:

1) International Collaboration: Fostering collaborative global partnerships to develop and refine trustworthy AI frameworks.

2) Continual Enhancement: Commitment to perpetual evaluation and refinement of these frameworks in tandem with the evolving Gen AI technology.

3) Public Education: Elevating public awareness concerning the potential advantages and risks of Generative AI, enabling informed decision-making.

Embracing these considerations, organisations can stride confidently into the future, propelled by the remarkable potential of Generative AI while safeguarding ethical integrity and societal well-being.

About authors: Sathish Gopalaiah is the president, consulting at Deloitte South Asia, and Payal Agarwal is partner, consulting at Deloitte India.

Also Watch:

The thoughts and opinions shared here are of the author.