Biggest Hurdles AI Industry Faced Amid The Pandemic

The Coronavirus pandemic virtually affected all sectors and forced everyone to work from home, leading to a number of challenges in everyday business processes. The AI industry was no exception. All aspects that go in the creation of AI, including the data, talent, investment, infrastructure, and research got impacted.

This article collates a list of some of the biggest hurdles the AI industry faced amid the pandemic.

The shift in emerging tech investment

While it is true that most of the industries suffered due to lockdowns and businesses went stagnant, resulting in lack of investment, emerging technologies became one of the last priorities for many firms. 

According to a KPMG report that surveyed 900 executives from Global 2000 companies, around four in ten people said that they would stop investments in emerging technologies, including AI during the pandemic. Around 57% of the executives also said that COVID-19 significantly changed their priorities. 

‘Weird’ training data due to unusual activity

When starting to build AI, one of the main things to consider is data. The quantity of data generated has not been an issue in the pandemic since transactions in most sectors got digitised, and in fact, it was more than usual. 

However, the data that is used to train algorithms is many times generated through our behaviour and the way we interact with various platforms. This changed dramatically after the pandemic — data generated through our ‘weird’ behaviour confused many current algorithms. Machine learning models are designed to react to changes but perform badly when the data differs too much.

This also affected the scalability of AI models. For instance, when people start to order too much of one product, the AI models which predict inventory requirements, break down as they are not equipped to cope with such demand.

Remote Managing of Data Security

The pandemic forced most people to work from home, many for the first time. While many people have the necessary equipment to work from home, there is a lack of standardisation in terms of security and connectivity. Sharing a common internet with personal computers also makes a well-encrypted machine vulnerable to cyber threats.

This issue becomes even more relevant for the AI industry as huge sets of data form the basis of their products, and ensuring data security, especially for personal data that is of the utmost importance.

Not working within the firewalls of the company increased the number of cyberattacks. As a matter of fact, around 66% of the Indian organisations faced at least one data breach or cyber-security incident while working from home. 

Upskilling in the age of AI amid the pandemic

As technology gets adapted at a rapid pace, the significance of upskilling and reskilling of employees becomes ever so prominent. There is a continuous need for the employees, especially in AI, to be up-to-date with the emergence of the latest innovative automation tools and research.

With the lockdown and everyone working from home, given the scale at which these reskilling activities are carried out, the pandemic presented a logistical challenge, especially for the big companies.

Collaborating from remote locations

Virtual collaboration is a lot different than in-person while actually sitting in the office and working together, especially when the AI application being developed requires teamwork. This affected the speed of work since communication between teams helps resolve errors quickly and gets results faster.

Also, a lot of knowledge that is usually shared through informal communication does not take place since people working from home set meetings only for well-defined agendas.

Wrapping Up

While there are so many challenges that the industry of artificial Intelligence faced amid the pandemic, AI remained true to its ‘intelligent’ nature. 

There were (some of them still are) many hurdles that the AI industry faced with the onset of pandemic and a sudden change in the work ecosystem. However, most of these challenges, like data security, reskilling activities, and collaborations, are now being resolved by developing AI solutions themselves.