Pandemic has hastened AI adoption, IBM executive says

About 43% of study respondents say their businesses relied more on artificial intelligence due to the health crisis

Seth Dobrin, IBM’s global chief AI officer, says businesses need a clear explanation of what technology regulations mean. Antonie Robertson / The National

The Covid-19 pandemic has accelerated the adoption of artificial intelligence technology after businesses were prompted to quickly digitise to adapt to the new normal, according to IBM’s global chief AI officer.

“Before the pandemic, we saw a lot of AI developing but most of it was not being adopted … [the] market was growing as people were investing in AI even though they were not getting real value out of it,” Seth Dobrin told The National in an exclusive interview.

"Overall growth of AI was relatively flat throughout the pandemic ... however, we saw a tremendous adoption that is a big change from before [pre-Covid levels] … we expect this rapid adoption to continue.”

About 43 per cent of executives surveyed by IBM said their businesses hastened the adoption of AI due to the health crisis.

Larger companies are about 70 per cent more likely than smaller companies to have actively used AI in their business operations.

More than a third of those polled by IBM said that making employees more productive (38 per cent) and finding an efficient way to interact with customers (36 per cent) influenced their decision to use the technology during the pandemic.

However, a lack of skills and growing data complexity were among the top challenges hindering the adoption of AI, businesses said.

The investors recognised the value of AI in two key areas: in enabling the rapid development of Covid vaccines and in finding breakthroughs in treatments using data science and machine learning, Mr Dobrin said.

However, the pandemic has also led to the need to have “common global laws” to extract more value out of the technology.

“Targeted regulations” are required to ensure companies use AI in a more productive manner, Mr Dobrin said.

“Having common global laws [is] always valuable. As a company, we operate in more than one country and different laws make it difficult for us to operate in a globally connected manner.”

“We are too big as a company and we can’t have 50 different ways to operate in different markets.”

In April, the EU proposed AI regulation that requires developers and users to abide by certain data management rules, record-keeping and transparency.

The proposal is a direct challenge to a commonly held view in Silicon Valley – that laws should not interfere with emerging technology.

Mr Dobrin said that the industry is “heading in the right direction".

“We also need a crisp and clear explanation of what regulations mean … that will be the core fundamental to the success of AI.”

Trusted and explainable AI is crucial to widespread adoption of the technology and the success of businesses, including how they maintain brand integrity and comply with regulations, IBM said.

More than eight in 10 businesses told the survey that the ability to explain how their AI arrived at a decision was important to their business. More than three quarters of them admitted that trusting AI’s output was fair, safe and reliable.

“Trust is a big issue that we must be worried about. It should be the foremost thing in mind, before beginning to implement AI,” said Mr Dobrin.

“Also, data should be owned only by the individuals … businesses should ensure adequate protection to their employees and customers so they know their data is being controlled by them in a way they think is appropriate.”

There is also a lot of business value in respecting and protecting people’s privacy and data, said Mr Dobrin.

“As a consumer, I have a choice to opt in or opt out of companies or apps using my data … they have to tell me why they are using it and provide me [with] the exact value in terms of experience. I should be comfortable and not feel [that they are] intrusive or creepy,” he said.

“End consumers should always know that AI is fair and how it reaches any decision. We also do not want gender biases or any socially unacceptable biases in AI-based decisions.”

IBM, which is working with Smart Dubai and the Saudi Data and Artificial Intelligence Authority on various AI-led initiatives, aims to have similar partnerships with various entities and ministries across the Middle East.

“We are working with Saudi Arabia and the UAE to develop AI and facilitate the transformation of their economies towards a more digital and tech-savvy ecosystem,” said Mr Dobrin.

To improve its AI capabilities, the company has invested heavily in natural language processing that is essential for businesses to communicate with customers, said Mr Dobrin, without disclosing the value of the deal.

IBM is also developing neuro-symbolic learning that will enable AI to mimic the human brain and explore the effects of quantum computing on AI or how exactly their combination will add value in solving intractable problems in the future, he said.