- Our Work
- The Reformer Blog
24 November 2016
Artificial Intelligence (AI) is a field of computer science and a set of technologies inspired by the way people use their cognitive functions to “sense, learn, reason, and take action”. Currently, AI technologies are tailored to particular tasks. Speech recognition technologies, for example, only know how to perform that specific task and none other. Despite, these technologies remaining mono-task in the foreseeable future, the current and potential impact of AI on daily life is huge. The breadth of its applications is in perpetual expansion. AI technologies can for example reduce energy consumption, lip read better than humans, build rationales for the decisions they make, help paralysed patient communicate through a speech computer by only using their minds and can be used to custom make products.
The rate of recent advances in AI continuously surprises the scientific community, despite it having been “patchy” in the past. Google DeepMind’s breakthrough earlier this year, with its AI technology (AlphaGo) beating the world Champion of Go, was 10 years ahead of what scientists thought was possible. Go is a complex ancient Chinese game which has more possible positions than there are atoms in the universe. To win human players often resort to instinct, which explains the significance of this breakthrough – the AI would have had to exhibit some instinct to beat the world champion.
If this is already the reality, what lies ahead? More specifically, what lies ahead for the applications of AI in public services in the next 15 to 20 years? As was mentioned in the previous blog of this series, the applications of AI in public services are currently nascent and the realm of possibilities infinite.
One of the more obvious impacts of AI technologies will be increased accuracy and efficiency of public services. A report by the Stanford School of Engineering, suggests that AI advances, in combination with sufficient amounts of data, will have a profound impact on the on the way healthcare works. Currently, when a patient goes to the doctor and describes his/her symptoms, the doctor tries to recognise the patterns of a known disease or condition. By 2030, AI will overhaul this process as doctors become supervisors of what patient information is fed into the algorithm and evaluators of the diagnostic output. The precision of diagnostics will increase and the probability of misdiagnosing will decrease. Healthcare robotics will also profoundly change the way care is provided, as their use increases.
AI technologies will also allow for the tailoring of policy programmes to suit individual needs, which will increase both their efficiency and effectiveness. According to the UCL Knowledge Lab the applications of AI in education have the potential to address the attainment gap. This is because AI technologies can tailor services at scale. In education, adaptive learning will become widespread. This technology provides individual support to students by adapting its learning content and techniques to the “student’s interactions in real-time”. The increased tailoring of services will have a profound effect on the achievement of better social outcomes. The JobCentre Plus of the future might very well use an algorithm that could individually tailor programmes for job seekers based on real-time data of available jobs, the person’s qualifications and skills and evidence about which training programmes are most effective to get that specific type of person back into work.
The potential to radically improve public services through the use of AI is substantial. The tailoring of services to fit the needs of the population will not only allow for the public sector to become more efficient, it will also dramatically improve social outcomes. However, the current structures to encourage its use in the public sector are lacking. In addition, safeguards need to be put in place to soften the blow of AI’s potential negative impacts. These will be the subject of the next blogs in this series.
Eleonora Harwich, Researcher, Reform