by Andrew Cochran
updated October 2019
A QUESTION COMES TO MIND, we say it out loud, an ‘assistant’ answers back. We glance at our mobile, it unlocks. We want another language, select from a menu, the translation appears. AI systems are ‘writing’ news, flagging news, and readying news — transcribing texts and researching files and suggesting copy and proposing headlines, all in seconds or less. Slowly but surely, AI is changing our relationship with time.
It’s not even slowly anymore. Artificial intelligence as a concept has been around for six decades. The noticeable impacts have been in the past six years. AI operates with high-performance computing coached by algorithms and nourished by data. Improvements spur the algorithms to weigh more predictions more quickly. One example: in November 2018, visual recognition in training sets was 16 times faster than 17 months earlier.
It’s easy to be caught up in the wave, but there are shoals on these shores. Capabilities for good work in reverse. AI systems have become unbelievably good at producing believable images. Variants of ‘deep learning’ can yield ‘
Work is underway to have AIs pick-out fakes, especially those generated by other AIs. It could lead to a kind of truth arms war. There will come a time when traditional verification, by seeing for ourselves, will no longer be conceivable. You can’t look a news-seeking algorithm in the eye. We’ll need to become comfortable deferring authority to AI systems like we came to accept results from pocket calculators. As with time, AI could change our relationship with trust.
Delivering a trusted report in a timely way is a bedrock of journalism. The last disruption, still called ‘the crisis,’ was essentially about business models. How could journalism sustain itself with revenues dropping against fixed costs? The next disruption will be about journalism’s soul: where is the locus of authority? Does it pivot on findings by humans or computers? On what basis should either be trusted?
And, will big data produce bigger journalism? Some newsrooms have been discovering insights from data at scales and speeds previously unfathomable for journalists. New ways of understanding our world may become possible, forming the foundations of a journalistic renaissance. Or will ‘real’ journalism become smaller, perhaps more intimate, with AI systems commoditizing breaking news and making algorithmically-customized storytelling more commonplace.
Either way, there will be an expectation for even more rapid delivery. It will be measured by an audience living in AI time. In comparison, ‘the crisis’ could seem quaint.
SEE ALSO