Myths, clichés and fantasies about AI

Everyone is talking about artificial intelligence, some with euphoria and others with panic. Some approach it analytically, while others speak with a drink in hand. Amid the noise, several predictions are repeated so often that they begin to look like certainties. The column says they deserve scrutiny.

The first myth is that adoption will be rapid. Individual use may grow exponentially, but institutional transformation moves at administrative speed. AI requires changes to processes, responsibilities, information systems, quality standards, budgets and power relationships. Testing it is quick. Integrating it is slow.

The claim that AI will eliminate jobs also needs qualification. In the short term, it may eliminate more tasks than entire positions. Many professions could lose autonomy, prestige and decision-making power. Before mass unemployment, the column expects some occupations to be degraded and reshaped. It also identifies a problem at the starting level: the tasks being automated often gave young people a way to learn a profession. Those routes into careers such as law and journalism are disappearing.

Nor will AI generate immediate savings. It initially increases spending on licenses, computing, consulting, training, cybersecurity, oversight and integration. Companies may keep paying for old infrastructure while adding new systems. “Before becoming a machine for saving money, AI will be a machine for spending and investment,” the article says.

More output is not automatically more productivity. Producing presentations, emails, videos and documents at almost no cost can shift the burden to the recipient, who must read, verify, select or ignore them. AI may make the producer more productive while making the receiver less productive.

The same distinction applies to talent and competitiveness. AI can broaden access to certain capabilities, but not guarantee good results. When everyone has the same tools, reputation, capital, judgment and willingness to take risks become more important. Companies will not become competitive simply by acquiring more technology. They need clear processes, reliable data and defined responsibilities. Otherwise, they may only automate disorder. Technology alone does not transform reality. What a machine can do and what a society ultimately does with it remain separated by a gap the article leaves open.

“

Before becoming a machine for saving money, AI will be a machine for spending and investment.