This week, PlanLab's Business Advisor, Michael Wirth talks about the difference between AI and Machine Learning and the strategic questions to consider when using this tech. Michael also shares some practical applications and his hopes of how charities will innovate in these areas.
1. How would you describe AI and Machine Learning in a couple of sentences?
Artificial Intelligence (AI) is the simulation of human intelligence by machines i.e. computers and its application, for example in voice recognition systems like Amazon's Alexa, computer games such as Google DeepMind's AlphaGo, or automated customer service using Natural Language Processing.
To equip machines with intelligence they need to be trained. A core approach to this is Machine Learning (ML) which exposes an algorithm to large sets of data from which it can learn and which then enables it to predict behavior based on the patterns "learned".
2. What are the main differences between the two?
Artificial Intelligence, Machine Learning, as well as Deep Learning are terms often used interchangeably and have evolved chronologically in that sequence since the 1950s. Deep Learning is a sub-set of ML which uses some ML techniques to solve real-world problems. It does this by tapping into neural networks that simulate human decision making by imitating the brain's connectivity, classifying data and detecting correlations.
Deep Learning requires huge data sets and computing power which have become possible and more affordable with recent advances in hardware and cloud capabilities. The more data that is available, the more accurate recognition and predictions will be, even when addressing complex patterns which require more detailed, specialised insights, for example with medical or geographical imaging data.
3. Where have you seen the potential for both?
The non-profit realm is commonly resource constrained, both in terms of money and people. Automation technology can assist in scaling and geographically expanding services that charities may employ or offer. Generally, reducing costs, maximising outcomes and optimising donor satisfaction will be the main benefits—considering that AI can augment human capability rather than necessarily always replacing it.
AI offers the tools to provide autonomous 24/7, global, real-time services, for example by giving donors tailored donation advice after office hours through chat bots (computer programmes that simulate human conversation through Natural Language Processing). Similarly, AI can make information available to beneficiaries in more accessible, customised ways in response to an individual's needs, for example automated crisis help lines.
Charities with international reach can break language barriers and for example employ real-time translation technology using popular applications like Skype that facilitate live conversations with refugees. This has the added advantage of protecting sensitive personal data which otherwise depends on interpreters' confidentiality. Similarly, translating promotional materials for global audiences, can be enabled by AI.
Some charities, for example those engaged in scientific research, can benefit from efficiently processing larger volumes of academic data than ever before . This will be a game-changer in areas like healthcare.
4. What are the strategic questions that arise when considering how to use these technologies?
Like any business, investments in the non-profit sector need to be weighed up against their return so, depending on the core activity of the charity, the value of AI should be assessed in the same way. Ask, "how can AI help us meet our organisation's mission and goals?"
Furthermore, tech industry experts are discussing the implications of AI in ethical, legal and societal terms for the areas of fairness; reliability; privacy and security; inclusiveness; transparency and accountability (see "The Future computed" and "Artificial Intelligence and Robotics"), answers to many of these are in flux due to the quickly developing nature of technology innovation.
For forward-thinking charities, in order to harness innovative technologies, they should consider appointing individuals with technology expertise in their organisations, on their boards, or even a technology advisory board.
5. What are your personal hopes for charities adopting these technologies?
The quality of AI benefits from access to data at scale which is costly. If the "third sector" were able to establish shared platforms to consolidate their intelligence from deep insight, this could make AI more affordable and accessible, benefiting a range of charities. This would be similar to shared platforms used by, for example, religious or humanitarian organisations that offer Customer Relationship Management (CRM) or Enterprise Resource Planning (ERP) capabilities.
AI-based automation tools enable charities to expand reach and increase the quantity and quality of direct interactions they can handle with limited human resources. This will accelerate the nurturing of supporters from first touch-point to active engagement, raising the level of involvement at every stage. Equally, employing automated, intelligent solutions will enhance the work non-profits engage in, benefitting the disadvantaged, the persecuted, the sick, etc. What better use for purposeful technology?
- A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence, http://raysolomonoff.com/dartmouth/boxa/dart564props.pdf, August 1955
- Mastering the game of Go with deep neural networks and tree search, https://research.google.com/pubs/pub44806.html, Nature, vol. 529 (2016), pp. 484-503
- AI is already having an impact on charity, https://www.cafonline.org/about-us/blog-home/giving-thought/the-future-of-doing-good/5-ways-ai-is-already-havin-an-impact-on-charity, June 2017
- The Future Computed: Artificial Intelligence and its role in society, https://msblob.blob.core.windows.net/ncmedia/2018/01/The-Future-Computed.pdf, January 2018
- Artificial Intelligence and Robotics, http://hamlyn.doc.ic.ac.uk/uk-ras/sites/default/files/UK_RAS_wp_AI_web_retina.pdf, 2017
Michael Wirth: PlanLab's Advisor