Why we need to get AI right from the start

The chair of the APPG on artificial intelligence says that new technology should be matched by progressive policymaking.

Sign Up

Get the New Statesman's Morning Call email.

Artificial intelligence technologies have taken on a central importance in our everyday lives during the Covid-19 pandemic. Not only have they been at the heart of public health strategies and pharmaceutical research in the fight against Covid-19, but data-driven tools were also essential to maintaining our professional and private relationships during the lockdown.

Many proponents of new technologies argue that AI has been the main force in creating a “new normal” with the potential to replace the life as we knew it pre-pandemic. However, one could equally argue that it is due to AI that our new normality is in most parts still reminiscent of the lives we lived six months ago. AI-driven applications made it possible to preserve a semblance of normality and routine that helped many of us through months of social distancing and limited access to stores and health services.

Whether the pandemic will fast-track the Fourth Industrial Revolution remains to be seen. However, our dependence on AI technologies in almost every aspect of our lives has been made abundantly clear during the pandemic. Nevertheless, with our increasing reliance on AI-driven applications to manage our day-to-day life, also their challenges and limitations surfaced in many unexpected ways. It became evident what impact skewed data, mismatched algorithms, and technological glitches could have on important areas of our private and professional lives.

Human interactions that were traditionally built on relationships of mutual trust had to be moved into the comparatively anonymous virtual realm. Video calls did not only create, at times, an unnatural distance that was augmented by the delayed response of our conversation partners, but which also deprived us of many of the nonverbal cues that facilitate social interactions. This has been especially challenging in matters that concern health and education, fields in which we value human expertise and dialogue especially highly.

Even though both sectors had profited immensely from AI technologies long before the current pandemic, the sudden reliance on online education and health services has fuelled the discussion in how far we can allow algorithmic decision-making into areas pertinent to our physical and psychological wellbeing. It might be that the pandemic only served as a catalyst that made us realise how interwoven our lives had already become with AI technologies prior to the onset of the global health crisis. However, under these new circumstances the question for the responsible deployment of AI received new urgency.

Public health and medical research are fields in which AI technologies have played a pivotal role in recent months. Whereas the discovery of new drugs and diagnostics had been advanced by AI long before the emergence of Covid-19, the contributions that AI technologies made to the speedy determination of the properties and behaviour of the virus and to the discovery of pharmaceutical measures to combat it cannot be overestimated.

AI-based applications have further been helpful in public health measures to help contain the virus. In a recent evidence meeting of the APPG on AI, Professor Tim Spector from King’s College London gave evidence of the impressive results of the Covid Symptom Study which was conducted with the help of a symptom tracker application. The successful deployment of this app offers valuable insights for AI-driven methods in public health that will have a significant impact on epidemiological research and the medical community beyond the pandemic.

Many of the most impactful AI-supported methods in the fight against the Covid-19 crisis have profited immensely from the international and interdisciplinary collaboration of experts. Setting structures in place that facilitate these exchanges in the future will be of great benefit to the advancement of more data-driven medical research.

AI-driven apps and digital platforms have also facilitated the sudden shift to online classrooms and e-learning. Applications that enable personalised learning also promise to provide students with support tailored to their individual learning needs. However, to fully reap the benefits of EdTech, a comprehensive introduction to these new tools is necessary.

Many teachers, parents and students faced the challenge of familiarising themselves with new technologies and ways of learning within days as schools shut. During this process, structural inequalities became painfully apparent in the way families were able to realise a smooth shift to online learning for their children.

Technical equipment, the provision of an adequate learning space for the child under lockdown conditions as well as parents’ understanding of the EdTech that was entering their home played an important role in the adoption of online education. For this reason, a sound education of AI that reaches all citizens will be indispensable for successful social inclusion and for making sure that every citizen can benefit from these new technologies.

UK businesses have also undoubtedly profited immensely from the introduction of AI technologies, even before the pandemic, as many processes could be made more accurate and reliable with AI-supported automation. However, the impact of Covid-19 has made the importance of a “human-in-the-loop” clear as predictive algorithms, that ran behind e-commerce and marketing failed when confronted with customer needs that did not correlate with the training data gathered before the health crisis.

The new demand that was created by the international health crisis called for experts to fix flawed AI systems that had been trained with now irrelevant data, which now needed to be updated to produce more reliable results.

As more and more businesses expand their AI strategy as a response to the crisis, the lack of in-house talent that is needed to deal with such unexpected events has become painfully obvious. For this reason, the government created a programme to increase the number of highly skilled workers in AI and data science, roles to which students with diverse professional and subject backgrounds can apply.

Above all, the pandemic has brought to the fore the urgent need for international and interdisciplinary collaboration to advance a fair and safe deployment of AI and the importance of making AI technologies relatable to all citizens to create a society in which everybody can benefit from technological advances. AI technologies offer great promises in countless areas of our life, but to realise this full potential, we need to get it right from the start.

Free trial CSS