Exploring the outcomes of hard to identify, inaccurate information being spread

Q-Series: Will Generative AI deliver a generational transformation?, considers that generative AI could lead to a surge in competitive intensity, transform the workplace and accelerate the pace of innovation. Its large-scale early adoption, coupled with the potential advancements in and benefits to productivity that have been identified, indicate that generative AI may become a general purpose technology for many people, in much the same way as computers, printers and the internet have.

There is a vast range of far-reaching, multifaceted ESG considerations related to generative AI. While opportunities exist, for example in education and medicine, we believe that some aspects and certain uses of the technology could be highly disruptive.

Generative AI models are known to be sometimes "confident and wrong” in aspects of their responses. This happens when a model "hallucinates" or "confabulates" pieces of information that are untrue or unfounded in all or part of its response. Inaccurate information can be spread easily by generative AI models using existing web-based misinformation. In our view, inaccurate information or "confabulations" generated by AI models and that relate to regulators, companies or public figures could be widely shared, with potential market implications.

Misrepresentation, including of a company’s social media accounts to post false/satirical commentary, could potentially impact share prices. The development and growing adoption of generative AI mean it's likely to become very difficult to discern real from (deep)fake. Deepfake videos and audio (when someone's likeness is convincingly replicated synthetically) could impact markets. We can envisage examples such as a (deep)faked video of a CEO issuing (false) news on a company, or a well-known investor appears on a ((deep)faked) video announcing a significant stake purchase or
sale.

In this Q-Series, it's estimated that generative AI could result in tens of millions of job losses. A significant structural shift in employment levels could put considerable strain on government services having to financially support the cohort of under- and unemployed people. As such, we view mass adoption of generative AI tools and the potential resulting structural job losses as a possible path to a paradigm shift, in that it may reopen the debate around a universal basic income. There has additionally been suggestion that generative AI tools should be taxed, so that the benefits of these technologies are shared.