A Special Report is content that is edited and produced by the Special Reports unit within The Irish Times Content Studio. It is supported by advertisers who may contribute to the report, but who do not have editorial control.

From cows to cancer: the many benefits of AI

AI is being used to improve productivity and efficiency in healthcare, agriculture, HR and talent management

Wearable devices like Fitbits are providing clinicians with data to help manage early-stage heart disease. Photograph: iStock

Wearable devices like Fitbits are providing clinicians with data to help manage early-stage heart disease. Photograph: iStock

 

Healthcare

Artificial intelligence is already being deployed in a variety of ways to improve healthcare and health systems around the world. Wearable devices like Fitbits are providing clinicians with data to help manage early-stage heart disease, powerful algorithms are utilising big data analytics to review millions of patient scans to identify emerging diseases at an early stage, while the same systems are being used to identify patients at risk of developing a certain condition.

One particular benefit of AI relates to advances in precision health, according to Microsoft head of public sector Frank O’Donnell. “This helps with better diagnostics,” he says. “Doctors will only treat what needs to be treated. Microsoft Genomics provides researchers and clinicians with cloud-powered genomic processing services to handle the heavy amounts of data they work with. We are working in partnership with St Jude Children’s Research Hospital in the US on a study to diagnose paediatric cancers.”

Another initiative undertaken by the company is a partnership with Seattle-based Adaptive Biotechnologies which is combining AI and machine learning with advanced biotechnology to create a practical technology for decoding the human immune system. The aim is to create a universal blood test that is capable of mapping a person’s immune system to detect a wide variety of diseases including infections, cancers and autoimmune disorders at their earliest stage, when they can be treated most effectively.

InnerEye is a particularly exciting project which can delineate cancerous tumours as well as healthy anatomy in order to enable clinicians to plan radiotherapy treatment and surgery much more accurately and efficiently. It uses state-of-the-art machine learning and computer vision to turn medical images into measurement devices to amplify a clinician’s ability to personalise treatment, spend more time with their patients, and save costs for hospitals.

“There is also great work going on to help clinicians gain rapid access to all the latest research literature on any disease they might be treating,” O’Donnell adds. “This will help them spend more time treating their patients and speed up accurate diagnosis.”

Also helping clinicians to spend more time with patients is Project EmpowerMD, a research collaboration between Microsoft and the University of Pittsburgh Medical Centre, which aims to create an AI-powered system that can capture and integrate medical expertise at scale by listening and observing doctors doing their work as they meet their patients.

The use of bots as intelligent care assistants has the capacity to improve health services. “We are providing hospitals with bots which can understand what patients are in different wards, how long they have been there, and for what reason. This can improve bed availability and support more efficient discharge,” O’Donnell says.

“This will become more prevalent. The technology will be able to learn about patients’ interactions with the systems, assist with their journeys through it, and support health services to become more efficient.”

The development of new drugs is also being assisted by AI. IBM’s Watson for Drug Discovery uncovers connections and relationships among genes, drugs, diseases and other entities by analysing multiple sets of life sciences data. Researchers can use this to assist with evidence-backed predictions, thereby finding the most promising new drug candidates, potentially taking years and hundreds of millions of euro off the R&D process.

Agriculture

The use of AI in the agriculture sector is advancing rapidly, according to Teagasc research officer Laurence Shalloo. “In the past, it was used mainly in tillage farming, but we are now seeing big advances in the livestock sector,” he says. “If you look back, farmers had technology like robotic milkers and so on and it was working well. Where it fell down was the interaction with the farmer.”

He explains that while the technology gathered and collected large volumes of data in order to support decisions, this wasn’t enough. “The farmer doesn’t necessarily want the data, they want the device to make the decision for them.”

He foresees a time in the not-too-distant future when automated milking parlours will be able to make quite advanced decisions for farmers. This might see it excluding cows with high cell counts in their milk to allow them to be treated for whatever is causing it.

“They could also detect when an animal is in heat, contact the artificial insemination service and recommend what straw to use to maximise genetic gain,” Shalloo adds.

“AI is already providing very practical solutions in the agri sector,” says KPMG partner in management consulting Owen Lewis. “For example, solar-powered robotics that are controlled by a mobile app can be used to perform tasks such as spreading herbicide or weeding large areas of land. Using GPS and sensors, machines can negotiate terrain, avoiding damage to crops, and can deploy very targeted and precise applications of chemicals, reducing waste and harm to the environment.”

Microsoft national technology officer Kieran McCorry explains how AI can help improve agricultural productivity. “Agriculture has always been a tough business – the margins are low, Mother Nature is cruel sometimes, so it needs whatever assistance it can possibly get,” he says.

“We believe that by collecting lots of data about specific locations on a farm and using algorithms to help farmers get insight across the full system that they’re trying to operate on, they can make much more precise decisions about how to plant and treat different crops. You almost want to get it down to treating each individual plant like it’s special. If you treat them that way, you can see transformative impacts in yield across fields.”

Microsoft led an agricultural AI project in India in partnership with the state government in 2017. “That project was an attempt to use machine learning to make predictions on when sowing and harvesting should occur,”says McCorry. “The year that we ran the experiment we put guarantees in place to ensure farmers’ income wasn’t under threat and we told them not to plant their seeds until our application told them to. They listened and planted two weeks later than their neighbours – and this was based on historical data, weather patterns and many other things. At the end of the season, those that had participated in the programme got 30 per cent higher yield on average than those that did not. That’s just one data point and one year in the world, so we’re trying to expand that out. It shows the potential.”

That same technology can be used to assist Ireland’s grass-based production system for livestock, according to Shalloo. “There is a big win there for Ireland. Very few studies have been done on grass production internationally. We can use AI to take a big leap forward and Ireland can be a world leader in this area.”

HR and talent management

“AI can support human resources [HR]professionals across a range of areas including the two key areas of the HR lifecycle – talent attraction and talent retention,” says Microsoft cloud and AI business group lead Paul Shanahan. “On the talent attraction side, we just have to look at what companies like Opening.io are doing to innovate the hiring process.”

The careers market is incredibly buoyant at present and finding the right hire for any company can be a very difficult process, he notes. “Their technology allows companies to find the right candidate faster. Their AI allows for a contextual search across the CV database, ensuring that employers not only spend less time searching but are more likely to find matches based on potentially relevant skills, and how similar their CV looks to previously successful candidates. It automatically ranks CVs against others that have applied and even goes as far as predicting salary bands for potential candidates based on their skills. The value this can bring to not only the hiring organisation but also the candidate in making faster, more relevant decisions on potential hires is huge.”

KPMG partner Owen Lewis says people must be open to AI’s use in the workplace. “AI should aim to enhance human thought rather than replace it,” he says. “Considering it in this context opens the door to see the scale of the opportunities, rather than just the threats.”

But we also need to be realistic. “In saying that, there are jobs that will be replaced by AI, but the impact of this can be better managed if we are collectively and proactively planning for a future that is open to and includes AI,” Lewis says. “For HR and talent management, there are many opportunities to improve the workplace and the worker experience, rather than replace the worker. There are obvious AI applications on the administration side, particularly for larger companies, but there are also other opportunities for employers to create a more personalised employee experience using data to analyse and identify employees’ different skillsets and motivators and optimise the workplace accordingly.”

On the retention side, Paul Shanahan believes company culture, rewards and recognition, personal development and work-life balance are all things that can be enhanced with AI-empowered technologies.

“Microsoft’s Workplace Analytics as part of Office 365 can give employers and employees real-time insight into their working day and week,” he says. “It will proactively promote adjustments to improve diary management, email habits and collaboration amongst specific sets of employees and teams. It’s all done via AI and, on a personal level, has helped me in reducing the email overhead I put on myself and on others by prompting me to turn off the laptop or to think about who I’m putting on the email.”

He notes that companies like Globoforce are pioneering the rewards and recognition space and supporting organisations in building a more human approach to recognition culture.

“In very simple terms, their service allows employers set up a peer recognition platform across the company with employees encouraged to participate in acknowledging people who go above and beyond. Their AI allows them to ensure that recognition is promoted fairly across individuals and teams and it promotes engagement from all aspects of the organisation, from remote workers to senior management. Tools like these go a long way to reinforcing a positive company culture and enable organisations to see patterns in behaviour as they develop so they can choose to promote or adjust them depending on the preferred outcome.”