Massachusetts Institute of Technology’s Prof Max Tegmark, in his latest book– Life 3.0 – hypothesises that the first indication that artificial general intelligence (AGI) is a reality will come from the media sector, as it offers the best influence-financial returns equation to help AGI scale.
But in a Chatham House discussion involving some of the industry’s leading venture investors at the Global Corporate Venturing Symposium the view by some was that AGI – artificial intelligence that is at least equally as powerful as human intelligence, was about 30 years away.
And while there are plenty of reasons to doubt predictions in AI, as Rodney Brooks, Panasonic professor of robotics at Massachusetts Institute of Technology and founder of Rethink Robotics, wrote in a post: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
But the short-term effects from AI-related developments are already having an impact through work in areas such as natural language generation (NLG), machine or deep learning, and generative adversarial networks, and these developments were the focus for much of the discussion in the roundtable.
Before this discussion started in earnest, there was an observation on the evolution of market capitalisations of media companies against disruptive tech peers. The main media groups’ market caps have grown over the past two decades, but at a rate outpaced by some players from the broader tech industry, such as electronics manufacturer Apple, internet conglomerate Alphabet, social media company Facebook, software developer Microsoft, and Chinese groups, such as Tencent, Alibaba and Baidu, that 20 years ago were either not actively involved in the media space or did not exist.
With these tech disruptors seen as the most focused on AI, particularly through corporate venturing units, such as Google’s Gradient Ventures, will this further exaggerate the gap in value and prove a tipping point for media companies?
While participants noted the shift to digital marketing and the increasing importance of data analytics, much of the discussion revolved around the general paradigm shift driven by AI.
Some of the participants said that, in terms of content production, what had fundamentally changed in the past few decades was the evolution of better equipment and streams of data, while the journalistic job remained essentially the same. While this had spurred the rise of data-based journalism and allowed a certain degree of automation, no machine could currently write an article better than a human.
Some noted that one of the hottest issues was content distribution over the online channels, which were dominated by companies like Facebook and Alphabet. Among other things, participants appeared to agree that there were opportunities on the monetisation side in the distribution and in terms of reinventing subscription models.
On the quality of AI-generated media output, it was also expressed that it was largely contingent on the quality of data available, so companies like Google and Facebook with large amounts of data held a competitive advantage.
Other applications of AI mentioned during the discussion were in image recognition or image analysis – used by e-commerce platforms. It was noted that may AI companies were sector-agnostic, so media companies should try to use them to help analyse the large quantities of data. This way, media companies could streamline internal processes, which would allow them to reduce the need for data analysts and scientists.
One of the participants drew a distinction between AI and machine learning (ML). Participants pointed to examples of ML used to identify fake news, enabling users to test the veracity of a story.
ML technology is also used to identify what people are wearing in video content and matching fashion retailers that may offer similar clothing. It was also noted that ML and AI had potential applications in call centres, human resources and recruiting technology, real-time multi-language transcription, translations, localisation, brand safety and customer services. One of the participants cited an example of a recruiting chatbot, with which job seekers were willing to spend more than 30 minutes on average. There was also an example of the ability to hire a bot to design a logo for a company.
During the discussion, it became clear that not all media markets are alike and the applicability of AI and ML may greatly vary across geographies. In Asia, media tend to be highly regulated, which requires editors to ensure no important stakeholders are offended or that content is not adult entertainment in nature.
The end of the discussion revolved around interesting existing strategies for media investors. Participants spoke of the new app market and ecosystem around Tencent’s WeChat or Meituan Dianping’s so-called super apps as great examples of how media could be converged with other commercial activities.
However, it was noted that one of the biggest challenges in integrating media, entertainment and commerce, such as buying clothes worn by stars on TV shows, lay in the difference in margins – double digits in media and up to 10% in e-retail. They also highlighted the importance of creating a relationship with end users.
One participant noted that consolidation was happening in the sector and in order to be a winner, one must first strive for scale. Participants generally agreed on the importance of partnering and collaboration, recognising that with new business models through blockchain and opportunities for decentralisation, the next generation of winners might not the be same as the current tech giants.
It was generally agreed that while media and telecoms were the top industries disrupted by AI, at the moment the overall focus was placed on complementarity – simplifying and organising people’s jobs – rather than replacing human labour. However, the most astounding example given was from one company, which had automated the equivalent of 40,000 jobs. The company had, nonetheless, managed to provide its people with other jobs, as it had expanded significantly.
Participants also touched on the issue of ethical concerns in media investing in a world encroached by AI. One pointed out the difficulty of deciding whether to invest in gaming companies, as games played by children and teenagers may be addictive. On a broader scale, the more pressing apprehensions touched on whether it was really ethical to invest in automation, AI or ML if people may be deprived of their jobs at some point in the future.
Or as Henry Kissinger, sectary of state for former US presidents Richard Nixon and Gerald Ford, said in the June issue of Atlantic magazine: “Philosophically, intellectually – in every way – human society is unprepared for the rise of artificial intelligence.” Kissinger gave three areas of special concern – unintended results from AI, changing human thought processes and values and making inexplicable conclusions. Management consultant Venkatesh Rao in his Breaking Smart blog similarly worried that AI would mean “our world will start to seem increasingly non-anthropocentric”.
This will matter less if people are changed – transhumanism. Interestingly, at about the same time 60 years ago that the first wave of AI optimism was evident, psychologist Carl Jung published the Undiscovered Self and pointed out that knowledge was not the same as understanding, and “ultimately everything depends on the quality of the individual”.
My thanks to all the individual contributors to the discussion and GCV Analytics reporter Kaloyan Andonov for taking notes.
