2017 will come to be known as the year artificial intelligence (AI) took off. While the science isn’t new — at FICO, for instance, we’ve been building AI-based fraud detection models for 25 years — AI is now being increasingly deployed across a range of industries, especially in the financial sector.
[clickToTweet tweet=”How and why #AI will evolve in #2018″ quote=”How and why #AI will evolve in #2018″]
So what’s next for one of the hottest trends in tech? There are five ways AI will continue to expand and flourish over the next twelve months:
1. AI Will Combine with Blockchain
Moving forward, organisations will start to use blockchain technology, primarily associated with cryptocurrencies, to record “time chains of events”. In these “time chains” people and their interactions are allocated an encrypted identity in a sequential chain of ‘blocks’. Since all the confirmed and validated transaction blocks are linked and chained from the beginning of the chain to the most current block, this shared distributed, digital ledger becomes the single source of truth. As such, it allows audit trails of data usage in models, particularly in data permission rights.
For example, consider renting a holiday villa. In the future, you’ll be able to book the property using a micro-loan that approves you to stay for a set period, like two weeks. This micro-loan will have information about your previous holiday rentals and events attached to its chain, along with a codified history of the property’s previous visitors, occurrences, and maintenance. Throughout your stay, details about your habits and interactions – for example energy consumption, breakages etc. – will be automatically recorded on the blockchain. When you leave the villa and lock it, the rental is complete and auditable on the chain.
In 2018, blockchains will also be used to illuminate the web of relationships within specific events to generate further insight and in some instances highlight out-of-the-ordinary and potentially damaging activities. For example, think about average daily interactions with money, people, places and things. Most days follow a set routine but sometimes chains of events occur that have new meaning, perhaps indicating bust-out fraud, money laundering activity, even suicide.
2. AI Will Learn to Fight Back
Defensive AI – systems that seed their outputs with “faint signatures” to mislead, confuse or identify the attackers learning the AI system’s response – will be front and centre in 2018.
Although we’ve long talked about how AI and machine learning (ML) will help companies drive differentiation in competitive markets, this principle also applies in the criminal world, where attackers are already using malicious AI to circumvent the AI models companies have in place. This arms race, in which criminals arm themselves with “adversarial machine learning,” tops McAfee’s 2018 security forecast.
3. AI Will Have to Explain Itself
The need for Explainable AI is being driven by upcoming regulations, like the European Union’s General Data Protection Regulation (GDPR), which requires explanations for decisions based on scores, including those produced by AI and ML systems.
In the areas of fraud and credit risk decisioning, the challenge of explainability is a longstanding one. While it is being dealt with successfully in these areas, other industries have entire sets of proposed explanatory algorithms that are either right, ineffective or flat wrong. Under the GDPR, there are hefty penalties for inaccurate explanations – making it imperative that companies correctly explain the decisioning process of its AI and ML systems, every time.
4. AI Will Augment the Workplace
In 2018, AI will augment much more of the workplace, not just through better software, but also through increasingly better versions of ourselves. Whether it’s drawing the information together for us to be superhuman at investigation, data recall, or improving how we learn new topics, AI will enhance our ability to process new information, especially in the field of fraud detection.
The objective should not be to jump on the AI bandwagon. Instead, it is to further decrease human workloads and increase productivity in compliance departments. In the short-term, traditional approaches will not be replaced, but rather, enhanced with AI that will take detection to the next level. This will help financial institutions generate new alerts as previously unknown patterns are uncovered with AI.
5. Chatbots Will Get Better at Understanding – and Manipulating – Us
The chatbots of 2018 will rapidly become more sophisticated, which will help to dramatically reduce the costs of routine customer care activities while improving the overall customer experience. In the coming year, chatbots will be able to quickly understand human tone and speech content, and thus predict the conversational paths that are the highest in value and can fulfil different objectives.
However, on the dark side, this subtle “engagement” can turn into manipulation when AI learn the words that sway our attitude, actions and elicit large-scale reactions.
Dr Scott Zoldi is Chief Analytics Officer at FICO. He holds 35 patents related to artificial intelligence. Scott blogs at www.fico.com/blogs.