5 ways AI is about to transform evaluation
Artificial Intelligence (AI) creates unparalleled opportunities while simultaneously posing new risks for all organizations.
Not only is AI able to process and analyze vast amounts of data at speed but can perform step-by-step reasoning and take actions to achieve determined goals. In a brave new world of supercomputers, driverless cars and humanoid robots, we explore how AI is set to shape evaluation and transform organizations by unlocking their enormous value-creation potential.
1. Speed saves lives
For an organization working in emergency like WFP, the speed information flows to inform decision-making is everything. AI can help leverage existing evidence at critical response moments, generate insights, costly to produce up until now, and possesses the ability to create complex simulations helping to predict outcomes of various global scenarios. From crisis managers to programme designers to policy planners, the potential to support different decision-makers is huge. Delivering the right evidence to the right people at the right time has never been so close.
2. Increasing efficiency, removing bias, improving trust
AI technologies like Natural Language Processing (NLP) possess the power to analyze, categorize, tag, retrieve and translate texts across multiple languages, while interpreting and understanding sentiment in communications. For the WFP evaluation function, it is invaluable. AI can draw from a multitude of past evaluations, summarizing and surfacing trends on results, recurrent factors for success and shortcomings.
Responding to a growing demand for tailored evaluation evidence, the WFP Office of Evaluation (OEV) has initiated a project trailblazing NLP within WFP, partnering with other teams. AI promises to retrieve relevant information from virtually all existing evaluation reports in the blink of an eye, against a process now taking days. And by systematically reviewing documents, AI can enrich the content of evaluation products and eliminate the selective bias that inevitably comes with manual, human-led approaches.
As well as automatically answering targeted inquiries, AI will help prepare new evidence products from the selected, quality-assured documents fed into it.
3. Analyzing context, tracking progress
AI can accelerate the planning and formulation of evaluations, supporting analysis of an intervention’s context and helping to identify challenges and risks linked to a programme while developing appropriate mitigation strategies.
AI could track in real-time and report on programme results, enhancing transparency and accountability. Real-time access to evaluative evidence would reinforce the role of evaluation as a learning partner and support adaptive management.
4. Knowledge sharing, recommender systems
Among knowledge-sharing innovations, recommender systems can automatically disseminate evaluation findings and answer stakeholder queries while creating personalized knowledge feeds or recommendations for staff based on their roles, interests, preferences and behaviour. This would give direct access to information that is most relevant and useful. Not only will AI be able to answer user requests for evidence immediately but, like Netflix or Spotify, offer a selection (of evaluation evidence) suited to your area of interest.
Beyond chatbots, generative AI — the creation of text, image or other media — can be used to provide personalized responses to inquiries, generate speeches, articles or social media content, raising greater awareness about WFP work externally and improving internal communication between different divisions and offices.
5. Humans with robots
While AI holds tremendous transformative potential, it will not replace human expertise. Adopting AI must be approached with caution, with emphasis on finding the right balance between utilizing its ever-improving capabilities and maintaining human oversight. Human input is essential in setting goals, being the guarantor of ethical values, making sense of complex context and ensuring the accuracy of AI-generated content. To that end, education and effective management of AI systems are crucial to maximize benefits and minimize associated risks. Implementing technical safeguards and holistic governance strategies are a must as we start piloting generative AI. But with careful and responsible use, AI could be a game-changer in supporting humankind to build a better, more sustainable world.
AI technology has never been more accessible, more inviting to use and catch a glimpse of the future. Just as Google revolutionized access to information — from a novelty to a pervasive aspect of our lives, AI will soon shift from a curiosity to the commonplace.
