The possibilities for Artificial Intelligence (AI) in the post-COVID-19 workplace are virtually limitless. The challenges for the successful introduction of AI in businesses, however, are not likely to be the tech, but will center on trust, training and transparency.

The possibilities for Artificial Intelligence (AI) in the post-COVID-19 workplace are virtually limitless. The challenges for the successful introduction of AI in businesses, however, are not likely to be the tech, but will center on trust, training and transparency.

These were some of the key takeaways from a FU.SE virtual panel convened in May to explore the topic of “How the role of AI and digital technology is accelerated by the pandemic”. Panelists were: Alberto Giovanni Busetto, Group Head of Data & AI, The Adecco Group; Raffaele (Rafi) Dalla-Torre, Director of Data Science, Salesforce; Marguerita Lane, Labour Market Economist, OECD and Nadjia Yousif, Managing Director and Partner, BCG.

Photo by ThisisEngineering RAEng on Unshplash

The session explored a wide range of issues, including the importance of gathering and securely storying high-quality data, privacy, the ‘soft’ side of introducing AI and automation, including engaging with the workforce and the importance of companies maintaining a continuous focus on training, re-skilling and up-skilling to ensure employees’ skills keep pace with changes in the workplace.

The pandemic as a catalyst of AI

“What used to be desirable is now an existential requirement. The pandemic has been a catalyst for front-loading investment in technology,” according to Alberto Giovanni Busetto.

In addition, the new ways we are working in are producing masses of reliable data, according to Raffaele (Rafi) Dalla-Torre: ”Data is the new oil. And COVID-19 is creating huge amounts of data on how we are working.” Whereas before the pandemic, if a company wanted to assess, for example, how long people spent meeting each other, you would be able to see how long meeting rooms were booked. But that would miss informal meetings and interactions. Now, with all meetings – formal and informal – having to be conducted online, meetings data is very precise, complete and unbiased.

For Nadjia Yousif, understanding how the mix of AI and people work in practical terms is the secret to successful uptake of technology. As an example of this hybrid human/machine – or ‘bionic’ – workplace, she mentioned a fragrance company which employs ‘noses’ – people who are experts and who test new fragrances. It has also invested in the development of an algorithm that selects and samples a vast range of new fragrances which it would never have discovered without the technology. But the final decision on the shortlisted new fragrances still remains with ‘the noses’.

According to Marguerita Lane, when automation replaces labour, but the productivity gains are only incremental, jobs may disappear. If the productivity gains through automation are greater, giving a greater economic benefit, then new tasks and new jobs can be created and, coupled with training, the benefits will more likely be passed on to workers.

Some key themes covered by panelists included:

Even if companies have not yet invested in AI or machine learning, they should still prioritize the collection and storage of new data. This is because organizations won’t be able to use their pre-COVID-19 data to inform their business decisions for the future. This should also help them to avoid the difficulties related to data integration further down the line.

AI and data science aren’t off-the-shelf products that can be plugged onto data to create solutions. Rather they are a set of capabilities and a methodology to help solve objective problems scientifically, with the potential to provide valuable input to people who will then wrestle with the subjective choices where there are no ‘right or wrong answers’.

Introducing AI and automation into the workplace should be a leadership issue as much as a technology project. CEOs must equip their workforce to understand the vocabulary and methodology of AI and data science, communicate clearly what the changes mean and involve data scientists as an integral part of the wider company team.

Impact assessments showing how AI could change jobs and what sort of additional training is needed for employees are useful, provided this follows the latest scientific standards. Firms need to inform and involve employees as AI is introduced, using pilot projects to enable employees to experience working with AI.

Re-skilling, up-skilling, and training must be a constant and enduring focus for a company to ensure new ways of working are productive for all concerned and to position AI as a colleague. According to a report titled "Artificial Intelligence in the Workplace" published jointly by The Adecco Group Foundation and BCG, only four percent of workers think their jobs will be fully replaced by technology, but more employees fear the prospect of constant change and adaptation required to accommodate new technology.

Trust in the process – and in the integrity and transparent use of data – is a paramount success factor. As far as AI is concerned, the title ‘CEO’ is becoming the Chief Ethics Officer, especially in light of the sensitive nature of much data held by companies – for example, HR or customer information. One panelist suggested the strong ethics frameworks governing clinical trials should be a blueprint for AI projects, ‘but even more transparent than clinical trials’.

Employee engagement on ethics goes beyond communications. Pushing the use of AI and data science into questionable ethical areas has already resulted in employee push-back in some leading tech companies. In addition, the ability of employees to take collective action when dispersed and remotely working could be under threat and they must continue to be actively engaged. Therefore providing channels for exploring these issues internally is essential.

As a summary, to prepare for the post-COVID-19 period, companies are encouraged to:

  • Create, log, track and store data/build databases. Ensure the company is measuring things that matter, even if you don’t currently have AI = data is the data scientist’s raw material

  • Make an ongoing commitment to re-skill and up-skill employees as part of a culture of lifelong learning

  • Communicate and engage the workforce early and provide channels for employee input and feedback as part of the deployment process

  • Explain the methodology of data science and AI

  • Operate within a highly transparent ethical framework

Watch the full recording of the webcast here:


News and Research