AI Offers Huge Potential, But it Won't Happen Overnight

Artificial intelligence is the topic of choice at most large tech conferences these days, and terminal week's Fortune Brainstorm Tech was no exception. It was the focus of a number of sessions, both on the main phase—where topics included AI and cocky-driving vehicles—and in many of the conversations at various roundtables and meals at the conference.

While nearly everyone at these conferences talks upwardly the huge potential benefits of AI, getting it deployed and working correctly seems to be a difficult and oftentimes lengthy process replete with challenges.

For instance, at a breakfast roundtable covering what yous tin can practice with AI, a number of participants seemed both very hopeful every bit to the potential benefits of AI and equally clear on the bug that companies face as they try to deploy it today. Speakers at the breakfast included two representatives of companies with major customer-facing operations—eBay and OpenTable—and two others that focus on enterprise applications—Box and Oracle. Each described different sets of challenges.

EBay Master Strategy Officer Kris Miller talked about how AI tin help to advise products as part of "an end-to-end customer journey," every bit well as virtually the company'due south try to build a central database containing all the relevant information near a unmarried customer—which could be used to provide real-time customization and personalization. It'south hard to create this, Miller said, and they are withal working out latency issues, among other things.

Another goal is to arrive possible for customers to take a motion-picture show of something—a purse or pair of shoes, for example—and have the eBay app immediately show customers similar items for auction on the site. This involves ingesting a huge number of images, tagging over 1 billion items, and so grooming the AI on these images.

OpenTable CEO Christa Quarles talked almost how AI is helping the company add new criteria into its search ranking, resulting in improve search, and hopefully, additional sales.

The goal, Quarles said, is to create the "ultimate recommendation engine." The problem is that users have dissimilar needs at different times, so context is highly of import, as is recognizing both implicit and explicit signals.

Quarles besides talked about using Alexa or other voice assistants for "conversational commerce," though she said at nowadays Alexa isn't very skilful as a "browse experience."

On the enterprise side, Box Chief Product Officeholder Jeetu Patel said that although he'due south convinced that in the long term AI "volition fundamentally change how people will collaborate with content," he's afraid of over-hyping the technology, and cautioned that such change will accept time.

Patel said Box is interested in AI in three primary areas. I is the "Box Graph," which aims to understand the relationship between two pieces of content—content and people, or people to people—with the goal of feeding a search engine or improving security via anomaly detection. A second area would be "Box Skills," or essentially applying AI to content for things like automatic object detection inside documents to place the content of video, audio, and images. Finally, he said, AI tin likewise be used to ameliorate qualify leads.

Patel said it's of import to be clear that Box doesn't own this data, only rather the individual business organization customer does. He noted that with AI, "beta cycles take longer," in role because y'all demand to be very careful near unintended consequences and you need to make sure yous are non doing something incorrect, such as exposing sensitive information. Patel too said that compared to things like the large image databases used in consumer applications, enterprises only don't take as much training data, and preparation has to exist done "per tenant" (in other words, for each enterprise individually), then we need meliorate algorithms that crave less data. Patel also said organizations will demand a "chief ethics officer" to make sure the data is used properly.

Kyle York, General Manager, Business & Product Strategy for Oracle, who joined the firm equally part of the Dyn acquisition, noted that Oracle has added AI components to many of its applications, in areas such equally Enterprise Resource Planning (ERP), Human being Resources (60 minutes), and Customer Relations Management (CRM), and mentioned the company's contempo acquisition of DataScience.com. York also noted that only x to fifteen percentage of enterprise workloads accept moved to the deject, and he said Oracle aims to brand AI and machine learning safe and secure.

York said at that place is both platform data, which Oracle tin can utilise to amend its products, equally well equally customer information, which remains the customers', and said there are opportunities for improved "governance tooling"—helping enterprises understand what data is valuable, what information is risky, what data customers could object to you collecting, and more.

The two enterprise vendors later discussed the issues surrounding propriety data, and the related dilemmas with the "derived value" that can be gleaned from analysis of the combined data. And enterprises don't want a competitor to do good from their data, Patel said, adding, for example, that yous "can't have Coke learn on Pepsi data."

York agreed that proprietary information sets are a existent event, and said that while anonymized information sets can be put together, you take to motion carefully because these tin turn out to disrupt privacy and potentially be detrimental to future business concern models. He said Oracle is "trying to democratize information," through things such as its Internet Weather Map, which brings together lots of data sets, including amass and anonymized data. We're still in early days when information technology comes to information openness, in his view, and "a lot goes back to human context."

In other conversations I had at the prove, I heard some perspectives on other bug facing AI. Cliff Justice, who heads KPMG's Innovation & Enterprise Solutions practice, told me that cultural issues are in fact the biggest trouble for most companies when deploying AI. Justice noted that to properly implement today's systems, you need to first tag a lot of content; then, one time a model has been created, employees demand to accept that it volition make mistakes, and determine what the correct respond should have been; and finally, create a new model and repeat. Only this all depends on line employees understanding that the system won't be perfect, and committing to take the time to identify and implement corrections. This is a large alter from the kind of activities these employees are accustomed to, in many cases, and making the change is not like shooting fish in a barrel, Justice said.

Cocky-Driving Cars and the Larger Impact of AI

On the principal stage, speakers discussed cocky-driving cars and some of the larger issues facing AI.

Transportation Secretary Elaine L. Chao repeated the mantra that the department intends to be "tech neutral" and not option winners and losers, emphasizing that "rubber is ever paramount."

Asked when nosotros volition see self-driving cars, Chao responded that these will happen "a lot faster than what some people think but not every bit fast every bit others."

Chao noted that the department issued a roadmap for such vehicles last fall, simply equally things are moving faster than expected, new guidelines volition be issued subsequently this yr. In particular, she said, everyone is moving quickly to "Level 2" autonomy—in which a man still has to impact the steering cycle—and that the department only has one application for a waiver for a "Level 3" vehicle, from General Motors. A procedure for granting that waiver hasn't yet been adamant.

Chao noted that there is a perception problem; 74 per centum of Americans say they would exist uncomfortable getting into a self-driving car. She referenced an accident in which an autonomous vehicle powered past Uber killed a pedestrian in Arizona, which "showed how delicate public trust is," she said.

Chao said it'southward important that we practice not terminate up with a patchwork of state regulations, simply said she is unsure whether regulations should be assembled past the federal regime or by states working together under federal guidance.

Mike Capps, Diveplane Corp Mike Ableson, General Motors, & Marissa Mayer, Co-founder, Lumi Labs

Another session featured General Motors VP of Strategy Mike Abelson and Diveplane CEO Mike Capps, formerly of Epic Games. They discussed "what impact AI will have on humanity" in a conversation moderated by Marissa Mayer, co-founder of Lumi Labs and former Yahoo CEO.

"AI will impact everything," Abelson said, and volition be the biggest modify in technology in the final 100 years. But when Mayer compared it to fire, Capps countered that fire is highly regulated, and mentioned how "two idiots" acquired a fire that burned a nearby forest in Colorado.

AI will change how people interact with all sorts of devices, Abelson said, and voice interfaces "volition experience a lot more than like Star Expedition really quickly." Capps said he's more than afraid of the Twilight Zone. "A black box scares the hell out of me," he said, and to that end he is working on "understandable AI."