This blog post highlights key points from a discussion at a roundtable held at the Royal Society.

A large satellite dish

Machine learning – the form of artificial intelligence which allows machines to learn from data to make decisions or predictions – is already permeating our everyday experiences, as we’ve discussed previously on this blog.

Much has been written about the impact of machine learning on employment, often focusing on the potential effects of automation. These effects play out differently in different sectors, and for workers with different skills levels.

To explore the impact of machine learning on work in specific sectors, on 13 July the Royal Society held a roundtable with representatives from the legal sector, healthcare, engineering, banking and accountancy. We discussed what machine learning might mean for these professions, considering its implications for business practices, skills, and the place of the professions in society.

This blog post highlights key points from the discussion.

The role, and tasks, of the professions

The professions have traditionally conveyed technical expertise, and judgement on technical matters, to consumers who would otherwise not have access to this expertise. However, technologies like machine learning have enabled the rise of new players, institutions, and non-specialists carrying out this role, or elements of it.

For example, since launching in 2015, the DoNotPay chatbot has helped 160,000 people to successfully challenge parking tickets. The bot uses machine learning to ask questions that help determine if the ticket can be appealed, for example on the basis of signs that may not have been visible nearby, or extenuating circumstances.

Examples like this demonstrate that machine learning algorithms can perform certain tasks to a higher standard than humans. With the right data, and the right algorithms, such systems could be used to eliminate bias in decision-making, or to improve analyses by considering a broader range of datasets or inputs than a human could. Machine capabilities are even extending into areas where professionals have relied on their judgement to make decisions in uncertain situations. Whilst the first wave of artificial intelligence tried to programme machines to directly mimic the human thinking process, the second wave of artificial intelligence found that machines can better perform tasks by working in different ways to humans. For example machines can now tackle issues of judgement by resolving the underlying issue – uncertainty – in alternative ways.

That said, there are other tasks where the potential for machine learning is less clear. For example, many professions carry with them a set of professional ethics or standards. Could machines ever replace this ethical role?

Furthermore, the ‘human element’ is still key for many professions, for a range of reasons:

  • Customers may want face-to-face interaction for certain services.
  • Certain, particularly adversarial, environments might require a higher level of social intelligence.
  • A figurehead might be needed to lead a team or ‘rally the troops’.

Overall, using a ‘task mind-set’, rather than looking at entire jobs, can give nuanced insights into changes in professional employment, and the best ways for skills and training to adapt.

‘Thinking algorithmically’: the skills evolution

What does this mean for the future of the professions? Our roundtable identified three key areas to be addressed:

  1. Progression: At present, machine learning is being used to automate tasks that are often completed by junior professionals, such as the research tasks of junior lawyers. This might alleviate the burden of these somewhat-mundane tasks, freeing up time for other activities, but it also raises questions about how junior staff can progress to more senior positions without “learning the ropes” in the traditional way, and what action can be taken to avoid “pulling up the ladder”.
  2. Skills requirements: As machine learning capabilities advance, the skills required for the professions will evolve. Data literacy and an ability to ‘think algorithmically’, with a basic understanding of technologies like machine learning, could become core skills for newly-qualified professionals.
  3. New roles: As machines take on tasks currently carried out by humans, new roles will be created, reflecting the new skills required. We’ve already seen the elevation of Chief Technology Officers (CTOs), Chief Data Officers (CDOs) and Chief Information Security Officers (CISOs) to the boardroom, demonstrating the importance of technology in long-term strategy discussions for a range of organisations.

There are already examples of professional bodies taking forward-thinking positions on the transformative effects of these technologies. For example, The Law Society has recently published a report on The Future of Legal Services and the ACCA has just completed a two-year research project, Professional Accountants – the future.

Bringing the professions into the 21st century

The professions have adapted to technological change before; the ‘big bang’ technology boom in the 1980s dramatically changed the role of stockbrokers, promoting an evolution of this profession into portfolio managers, and other roles in the banking sector.

As for machine learning in the professions, questions still remain. While the role of machines might be suited to the technical, the extent to which they could replace the ethical side of this work is undetermined.

To reap the benefits of this technology, the professions may need to consider how they can create new kinds of ethics and flexible ways of thinking that allow roles to change and new business practices to emerge.

The Royal Society is currently carrying out a project on machine learning, which is considering the potential of this technology, and the challenges that come with it. You can read more about machine learning in this interview with Marcus du Sautoy.


  • Susannah Odell

    Susannah Odell