Opportunities and challenges AI will face in the coming year
Experts weigh in on what we can expect from AI in 2018.
What will happen in artificial intelligence (AI) in 2018? Four experts–Chad Meley and Stephen Brobst of Teradata Corporation, a leading provider of business and data analytics solutions, and Atif Kureishy and Eliano Marques of Think Big Analytics, a business outcome-led global analytics consultancy–share their predictions on the new technologies, business opportunities, and challenges the AI field will see in the coming year.
Chad Meley, vice president of marketing, Teradata Corporation
- AI infrastructure will improve dramatically. Survey respondents from 260 large enterprises told us their most significant barrier to realizing benefits from AI was “lack of IT infrastructure,” surpassing all other headwinds, such as little or no access to talent, lack of budget, and weak or unknown business cases. Poor software integration between various open source software components, along with GPU environments that do not meet enterprise service levels, have become top priorities to be addressed by vendors and the community. They will respond in 2018 with enterprise-grade AI product and support offerings that overcome the growing pains associated with new advanced technology. This will not only accelerate the value that comes from AI, but will also mark a shift in enterprise AI algorithms from running almost exclusively in the cloud (given its position as the path of least resistance), to a more balanced mix between public cloud and on-premises deployments that more accurately reflect the realities and foreseeable intentions of large enterprises.
- The first mainstream killer app for enterprise AI will happen in financial services. A large majority of the top 15 banks in the world will begin to share publicly their results in using AI to thwart bad actors across a variety of financial crimes, such as credit card fraud, synthetic identity theft, AML, and others. Unlike the traditional systems in place that are based on manually set rules and light analytics, AI-based systems using a wider spectrum of data along with lightly supervised algorithms will prove to be an order of magnitude more effective in detecting financial malfeasance with respect to accuracy, completeness, and timeliness. One could argue that deep learning is being applied in retail already for next-generation recommendation engines–while true, it’s limited to only the most advanced digital native retailers, and therefore excluded from this “mainstream” characterization. AI use cases from other industries such as health care, manufacturing, and transportation will progress in 2018 with some key one-off wins by select companies, but will not see a widespread break-out in 2018 to the degree we’ll see in financial services. Things are lined up perfectly for the killer AI app for fighting financial crimes: a well-understood and high-impact business case, recognition of the need to fight fire with fire (as the bad actors are using AI against the financial institutions), and a couple of innovative banks that have done it in 2017 and are openly sharing the blueprints with other banks (“the enemy of my enemy is my friend”).
- The need to develop a corporate AI strategy at the end of this decade will attain the same level of importance and urgency as developing a web strategy was at the end of the 1990s. Just as the late 90s saw the Web extend its impacts across a much broader swath of companies and governments than the preceding, formative years, so will the end of this decade be remembered as the time when every company began to define their vision and roadmap to embrace AI. Unlike other recent tech trends around big data and the cloud, where large enterprises could get away with waiting for the technology to work its way into daily operations, AI is more akin to the web in that first movers who get it right are positioned for a winner-take-all scenario. Wall Street will adjust valuations on evidence that large enterprises understand and appreciate the transformational impact of AI, and are making commensurate investments. Experimentation and pilots will not be sufficient to an anxious board and CEO demanding thoughtful and bold plans, placing some CIOs under immense pressure, whereas CIOs who embrace the transformation through well-communicated AI strategies marked by meted business outcomes along the way will thrive.
Stephen Brobst, chief technology officer, Teradata Corporation
- AI will pass the peak of inflated expectations. There will be a backlash against “AI hype” and more of a balance between applying deep learning and shallow learning to business opportunities.
- Concerns over explainability will deepen. Research into determining how a model treats its inputs and reaches its conclusions, in human-understandable terms, will make deep learning more acceptable for consumer-facing applications.
- GPUs will be widely adopted. General-purpose GPU technology will transition from niche uses cases in processing visualizations to widespread adoption for deep learning.
Atif Kureishy, VP, global emerging practices | AI & deep learning, Think Big Analytics
- AI developers will focus on being production-ready. After launching their AI initiatives, organizations will now shift their investment to operationalizing the lessons learned in AI. Some 80% of enterprises have invested in AI, which means that in 2018, we will see a focus on key challenges of workforce optimization (increasing machine-person ratios), and upgrading technology infrastructure (machine and deep learning platforms, GPUs) and analytic ops for rapid operationalization of learning models for machine intelligence. This operationalization will allow scaling (where runtime implementations of machine and deep learning models are able to address more requests as more users, concurrency, and workload are increased) and enterprise security (ensuring information confidentiality, information integrity, and the availability of the machine intelligence platform within production). Executives will launch more diverse and all-encompassing efforts to realize a more dramatic future after assessing the initial AI successes they’ve experienced. These efforts will have more risk in execution given the high-stakes perspective of a “winner-takes-all” scenario within the AI revolution.
- AI will spread throughout departments and rise to the C-suite. Beyond the bot, improved human-to-machine interactions will enhance AI for enterprises. Systems of insight–the enterprises’ capabilities, environments and platforms for data science and advanced analytics–will get a makeover to become more informed, contextual, and intuitive. In 2018, machines will sit side by side with executives and employees on key decisions and build a corpus of corporate intelligence that far exceeds existing enterprise efforts to capture and manage knowledge. The use of AI in both strategic and tactical decision-making will also improve efficiency, which will have a big impact on businesses like retail banking and telcos that are suffering from compressed profit margins.
- We’ll see more investment in trust-based reinforcement learning. In 2018, organizations will continue to be challenged by the vast undertaking of managing labeled data. From autonomous driving to fraud detection, supervised learning techniques require a large amount of labeled training data to increase accuracy of models. For some organizations, the task will be overwhelming and unfeasible, so investment in reinforcement neural network (RNN) technology will spike. RNNs use a reward and policy framework to learn based on a target response. Furthermore, there must be a proportional effort to build authentication and authorization into these AI models as they progress. As we push more intelligence into a machine that then collaborates with other machines, the environment needs to trust all its users. But, advancements in an adjacent field that seems poised to get just as hyped as AI could hold some answers: blockchain.
Eliano Marques, head of data science international, Think Big Analytics
- AI will reach low-latency. Scoring engines for complex deep learning algorithms will evolve to a state that they could run a production system with very low latency, a level of performance that today is difficult to achieve. This will enable more traditional use cases to be converted to deep learning and run in production at scale with low latency.
- Deep learning will reach larger areas of users. Keras for R and the TensorFlow ecosystem for R will make these new types of machine learning algorithms more widely understood and used. Millions of data scientists will adopt and rebuild machine learning pipelines on these tools. This is not to say there will be a shift from Python to R, but a convergence of Python and R users around the same solutions, as can be seen with packages like sparklyr (Spark for R programmers). What does this mean to an enterprise? More use cases will be sped up and improved with new techniques, new insights will be created, and more demand for new types of production-ready systems will emerge.
- AI will fuel a health care boom. Health care has seen several experiments with advanced AI-based analytics in 2017, often from Stanford, where new algorithms are beating doctors in diagnosing several critical diseases. Now it’s time to bring to real life machine diagnostics where AI starts to save lives.
This post is a collaboration between O’Reilly and Teradata. See our statement of editorial independence.