SINGAPORE: Generative artificial intelligence (Gen AI) promised a roaring business revolution, but 18 months after ChatGPT's big bang debut, its impact is more of a sob.
Observers say companies underestimated what it would take to leverage next-generation AI and overestimated its business impact.
Consultancy firm McKinsey predicted in July 2023 that artificial intelligence would add US$4.4 trillion annually to global corporate revenues, but in April, companies began piloting artificial intelligence and the technology That said, there are no more than a few use cases or situations where it applies. It can be used.
Gayatri Shenai, senior partner at the firm, said in a webinar: And this is where most organizations struggle. ”
The hubris following ChatGPT's debut in November 2022 may explain some of the disappointment.
Matthew Candy, global managing partner of Gen AI at IBM Consulting, said that some of the test projects that companies choose in 2023 are scalable, such as launching Gen AI for the ability to run with eight people. He said that the company would not have been able to provide any value.
“There was pressure, there were questions from above, and a lot of people felt like they had to do something or show something,” he said.
In a late 2023 poll of 300 executives conducted by MIT Technology Review Insights for Telstra International, companies that first adopted Gen AI reported less trust in their technology than other companies. I understand that.
Michael Schrage, a researcher at the MIT Sloan School, commented on the findings: Too many professionals see Gen AI as a way to automate or enhance existing workflows and processes, rather than rethinking the fundamentals of their use cases and what outputs and outcomes they really want. ”
Illusions – the Gen AI’s tendency to fabricate things based on probabilities – are a major obstacle.
OCBC Bank, which introduced the ChatGPT bot in October 2023, has 250 AI models and a 150-person data science and engineering team overseeing them.
Test the accuracy, limitations, and biases of large-scale language models using a number of tools.
Donald McDonald, head of the bank's group data office, acknowledged that “Gen AI has many businesses excited about its potential but are currently finding it difficult to deploy in production.”
Organizations also lack AI enablers, such as an AI strategy, people, data, infrastructure, operating model, and culture.
Geraldine Kor, Head of Global Enterprise at Telstra International, said businesses must have fit-for-purpose data architectures, sufficient computing power and robust connectivity to handle massive amounts of data. says there is a need.
Enough data to generate real-time insights is another biggest constraint.
A survey conducted by Salesforce and YouGov around March found that around 8 out of 10 Singaporean workers said data captured by AI models must be accurate, comprehensive and secure before they can trust and use it. I answered.
“In real-world applications, you can't rely on luck. You need to get people to trust the technology,” said Professor Lam Kwok Yang, a cybersecurity expert at Nanyang Technological University's School of Computing, on May 7. stated at the meeting.
Additionally, there is a shortage of AI talent, and short-term relief is unlikely.
A global poll of 2,000 executives from nine developed countries conducted by recruitment firm Adecco and Oxford Economics in late 2023 found that Singapore topped the list of employers planning to hire external AI talent. became.
Betul Jenk, Adecco’s head of Asean, was not surprised, given that 34% of expatriates have only been with the company for one to five years as of 2023. Employers may find it more economically viable to hire rather than reskill, even if it's not sustainable, she said.
She said: “This approach risks creating skills shortages that could drive up salaries and increase competition in the job market.”
Still, employers are too wrapped up in the AI race to focus on employee training, another study by workforce solutions provider Persol Kelly found.
In a February survey, just 27% of local employers said they had comprehensive plans to help employees replaced by Gen AI find new jobs, compared with the Asia-Pacific average of 36%.
“Despite being recognized as a company's most important asset, employees are frequently neglected,” the report said.
The question is: How many companies are ready to effectively implement Gen AI?
Workers will remain at the core.
“We're excited to be able to deliver a new experience to our customers,” said Brad Anderson, president of product, user experience, and engineering at experience management company Qualtrics. But when they think about how they want to use his AI, they want their customers and employees to be in the passenger seat assisting them, rather than having the AI replace the driver's seat. It is clear that it is. ”
He added that customers are using the company's AI tools to accelerate customer interactions with human agents and help managers act on both customer and staff feedback.
IBM's Candy said companies will realize they need a patchwork of models and technologies rather than a single AI generation model.
He said: “People are starting to realize the complexity of the landscape that is going to evolve, so how do we govern it, how do we manage it, how do we build structures and layers that can connect all of this? You need to think about different techniques and skills. ”
Companies that don't have enough data to use Gen AI may start in areas where data reliability is less important, suggests Sujith Abraham, Salesforce general manager for the ASEAN region.
“For example, account summaries and case summaries are not as dependent on the quality of the data. Companies can start there and continue to improve and refine the data through experimentation,” he said.
Anne Huitt, president of the Small Business Association, said many small and medium-sized enterprises are using next-generation AI tools for tasks that improve personal productivity rather than transform jobs or processes.
These companies could potentially integrate gen AI into their business processes, such as sentiment analysis and enhancing customer service with fast and accurate first-level customer inquiries.
“A 20% increase in productivity for an owner or team can have a significant impact on operations,” says Anne. — Straits Times/ANN
