We’ve studied the future of the internet since 2004. Here’s what we’ve learned.

Since 2004, the Pew Research Center has been asking thousands of experts to predict the impact of the internet in the future. In the summer of 2016, the Center and Elon University’s Imagining the Internet Center fielded the seventh such canvassing of technology analysts.

The latest batch of insights from these experts has been released in a series of reports exploring the future of fake news, trolling and free speech, jobs and jobs training, algorithms and the Internet of Things.

These predictions also inform a new article I wrote in the Pew Charitable Trust’s Trend magazine about how the pace of technological change poses enormous problems to economic, social and legal systems. I highlight four major changes tied to the ubiquity of technology, especially through the Internet of Things and the disruptions it causes. First, technological innovation will create a “datacosm” that infuses data into almost every nook and cranny of life. Second, algorithms will become more important in understanding and implementing insights from that plethora of data. Third, humans will continue to develop a new relationship with machines and complementary intelligence. And fourth, all this change will produce innovation in social norms, collective action, status credentialing and laws.

The Center has pursued these medium-term “futurism” efforts for more than a decade because those in our expert cohort have been prescient about major trends in the past. They have anticipated such things as the rise and potency of cyberattacks and extremist communities, the shift from desktop to mobile connectivity, the rise of (and challenges posed by) big data, the evolution of privacy concerns and shifting views about anonymity, the appeal of augmented and virtual reality and the baffling unpredictability of what technology will become “the next big thing” — even to those who invent the future for a living.

Looking ahead, these experts predict more disruption, starting with digital interfaces: They say voice-recognition (think “Siri” and “Alexa”) and “air typing” on connected surfaces of all kinds will become mainstream activities. A data layer will be superimposed on the physical environment, allowing more context, history, and insight to be gained when people interact with their surroundings. Artificial intelligence will be ever-more-deeply ingrained in decision-making and logistics. Gigabit-enabled applications will be common — as people interact with holograms of others in both artificial and real worlds that otherwise may not be easy to get to. They predict the very notion of “being present” with another person or encountering another place will change. In all, the expert respondents to Center questions argued that by the year 2025, the internet will become “like electricity” — less visible, yet more deeply embedded in people’s lives for both good and ill.

What do they mean by “ill”? They worry about the prospect of greater social and digital divides, including new divisions that arise as some people cope well in the new hyper-connected reality and others fall victim to its distractions and pathologies; that bad actors will exploit the new affordances of the internet to hurt others and worsen social intolerance; that nations (especially authoritarian ones) might take steps to balkanize the global, open internet; and that privacy will erode to the point where it might just become a “commodity” that only the well-positioned can purchase, while others cannot escape surveillance.

Perhaps most urgently, many experts worry that robots and artificial intelligence will displace existing workers, and possibly at a pace too fast for economic actors to create newer, better jobs to replace those that have vanished. They are skeptical, too, that government institutions and policy makers will act to mitigate the wrenching societal changes this would cause.

Yet, others are not as downcast about the future of jobs. In fact, our first expert canvassing about the future of jobs in the age of robots and artificial intelligence produced a split verdict: Roughly half of respondents predicted that robots and artificial intelligence will destroy more jobs than innovators can create; while half believe the opposite. The optimists point to history’s pattern where technological change produces serious economic stress, but eventually yields more and better jobs and a wealthier society that has higher levels of well-being.

So what’s next? We will soon ask about the future of truth-verification systems. We will question these analysts about how successful human ingenuity and machine insight will be in helping people find reliable information — or whether darker forces will prevail in tomorrow’s information wars. Once we get their answers, I will report back in this space. If you’d like to participate in this expert canvassing, email info@pewresearch.org with “Future of the Internet” in your subject line.