Understanding and Using AI Models

Artificial intelligence (AI) in its many new forms is being integrated into a wide area of our work, promising to change the way we work, make decisions and interact with the technology. Weak or narrow AI are good at performing specific tasks such as recognizing and responding to voice commands. This usage of AI is evident in image recognition software, self-driving cars and AI virtual assistants. We are yet to leverage on the full potential of AI that can learn, think and perform various actions similar to a human. A glimpse of it is already visible in the generative AI models such as GPT-4 from OpenAI and Google’s Gemini.

We can already see the evolution of large language models such as GPT-4. It can now understand image inputs and help generate and ideate on various creative and technical tasks. It’s a boon for businesses as it can help in the creation of targeted marketing advertisements. It is also being used in processing and analyzing data to extract valuable insights. These insights can provide an edge to businesses in terms of recommendation on tailored products, services and marketing campaigns.    

Several large businesses are also understanding the immense potential that businesses can tap through the development of AI. Top tech companies like Google, Amazon, Microsoft, NVIDIA and Salesforce have all invested the most in AI startups in 2023. Nearly all business owners believe that ChatGPT will help their business, over half relying on it for cybersecurity, fraud management and customer service. Over 60% of business owners anticipate that AI will not only help improve customer relationships, but also increase productivity. 

Globally, governments are realizing its potential use case in enhancing government services. Government Technology Agency of Singapore is exploring ways to integrate ChatGPT-4’s capabilities in providing various public services. In its 2024 budget statement, the government has announced plans of investing more than SG$1 billion in the next 5 years to develop skilled AI talent, robust AI infrastructure and enable industry digitalization in the country.

The Brazilian government is also advancing its AI ecosystem through the National Strategy for Artificial Intelligence (EBIA), launched in April 2021. This strategy aligns with OECD AI Principles, focusing on inclusive growth, human-centered values, transparency, robustness and accountability. The country will also host the Horasis Global Meeting in Vitória, the state capital of Espírito Santo, from October 25-26, 2024. This two-day event will gather leaders from business and government to address pressing global challenges, including climate change, inequality and peace.

Concerns Remain

Although AI is set to revolutionize the way we work, its extensive use also raises concerns for businesses. Businesses fear that their visibility on search engines will get hampered as more consumers plan to use large language models such as ChatGPT to search or gather information instead of search engines. Increased technology dependence is also a growing concern. Four of 10 business owners fear that we are becoming too reliant on AI, while 77% fear losing jobs due to AI advancements. AI-generated misinformation is also another growing concern, which if unchecked, can significantly impact brand reputation.

Lack of emotional intelligence is another significant barrier on AI’s path to human-like capabilities. Gaps in understanding or interpreting emotions can render large limitations in AI output, that may at times be insensitive or offensive.

AI skills shortage is another major hindrance in AI’s intended growth. A recent Salesforce survey reveals that 6 in 10 IT workers lack the needed AI skills in implementing AI, with IT professionals from the public sector being the most vulnerable.

Addressing Limitations

To address these limitations, there is growing interest in developing AI models tailored to specific domains. Domain-specific AI can be trained on specialized datasets, incorporating the deep knowledge and expertise required to excel in a particular field. Creating domain-specific AI involves not just training models on relevant data but also incorporating the insights and expertise of domain experts. This collaborative approach ensures that AI solutions are not only accurate but also aligned with the practical needs and challenges of the field. Moreover, continuous learning and feedback loops are essential to keep these models updated with the latest knowledge and practices.

Business leaders play a crucial role in driving AI adoption within their organizations. They are important in fostering a culture of lifelong learning and adaptability of AI among their employees. CEOs will also need to inculcate a culture of data that is unbiased and ethical. Putting together a team of data scientists, tech experts, operational managers and commercial leads under the leadership of a CIO or a CTO is also the right path in enabling a data-driven company-wide effort. 

Governments around the world are implementing regulations enabling sustainable development and deployment of AI that is beneficial for all. Regulators must work closely with AI developers, industry experts, and other stakeholders to establish standards and guidelines that promote the development of reliable and trustworthy AI solutions.

In this rapidly evolving landscape, the key to unlocking AI’s full potential will be to ensure that AI is not just powerful and versatile, but also deeply knowledgeable and specialized. Only then can AI truly deliver on its promise to revolutionize the way we work, make decisions and solve global problems.

Photo Caption: Harnessing AI’s potential responsibly and sustainably will be key