AI has been hot in 2016, and it’s not cooling off anytime soon. The investments, acquisitions, trials, reorganizations and breakthroughs of the past year have set the AI industry up to have tremendous impact over the next twelve months. Here are nine predictions to watch for in 2017:

1. We’ll end the hyperbole

AI will “land.” We’ll stop talking about far-fetched, man-versus-machine Skynet predictions and instead figure out how to harness AI to turn the slag pile of big data into the orderly summit of our dreams. The hype curve will calm down as people realize what AI can do and is doing, and thus form more realistic pictures of what it will do. We’ll embrace AI as critical for our economic productivity. From business process outsourcing to computer vision to natural language processing and more, we’ll strengthen today’s AIs, integrate them into our existing systems, and enrich lives and improve businesses as we go.

2. We’ll incorporate and develop AI broadly

AI will become pervasive in all software, business processes, and big data solutions. Disruption will come from small companies (e.g., Algorithmia, Clarifai, GumGum, Init.ai, Orbital Insights, Sentient Technologies) and large enterprises (e.g., Amazon, Apple, Baidu, Facebook, Google, IBM, Microsoft, Salesforce, Softbank, Tencent).

3. We’ll seek specialized human knowledge

We’ll have the broad epiphany that we need the right humans, in the right loops, to make AIs work with and for us. I’ve worked with world-leading experts at dozens of companies who are struggling to apply AI because of the scarcity of accurate, domain-specific human insights. After all, algorithms such as neural networks can learn without explicit instructions but need a tremendous amount of training data.

4. We’ll build ‘smart’ apps only

For example, shopping will be smart(er). We’ll see computer vision, coupled with the ability to intelligently crunch massive amounts of market and personal data, serve up better online shopping experiences (securely). As another example, today Siri, Cortana, Google Now, Alexa, and other natural language interfaces are exciting, but often generic, frustrating, and hilariously wrong. In 2017, we will use voice as a primary means of interacting with computers across a number of more useful, specific, reliable scenarios.

5. We’ll advance health care

It’ll be slow, but it’ll be in ways we desperately need. My personal dream is that we digitize the tremendous backlog of medical records in a way that helps medical professionals provide more intelligent care more efficiently, like our economy needs them to. My audacious dream is that the Department of Veterans Affairs leads the way.

6. We’ll advance agriculture

Drones will monitor and service crops from the air and ground, and AI-powered software will process the resulting data to help farmers make better decisions. After all, we have the “minds of hunter-gatherers, but our cuisine is that of ancient farmers.”

7. We’ll automate some jobs

Traditional jobs in transportation (e.g., trucking, taxis), plus warehouses and manufacturing, will get increasingly supplemented and replaced by robots operating in part based on AIs, with the aid of computer vision, autonomous navigation, and other forms of remote sensing and decision making. It’s a mixed bag. Jobs that used to be manual, dangerous, expensive, or impractical will become automated. But we need to proactively retrain people and create new jobs.

8. We’ll enhance security

AI will become vital to our national security. We don’t talk about security enough in startup-land, and I don’t just mean cybersecurity. We need AI to keep our borders, electrical grids, police officers, first responders, and “soft” infrastructure safe.

9. We’ll solidify ethics and guidelines

We need to do all of this ethically. Think of AI as an incredibly powerful scalpel. If we cut at the wrong angle, even slightly, the implications for society are profound. I hope that our new federal government 10X-es investment inside the Defense Advanced Research Projects Agency (DARPA), and publishes voluntary guidelines the industry can adopt to reduce bias in AIs, and that it does so after listening to industry leaders at large and small companies.

Maybe I’m feeling optimistic with the clean slate of a new year, but I believe AI is poised to revolutionize our lives in many ways. I for one can’t wait to see what 2017 brings to the industry. Would love to hear your thoughts, dreams, and fears, too.